X-Git-Url: https://git.whamcloud.com/?p=fs%2Flustre-release.git;a=blobdiff_plain;f=lustre%2Finclude%2Flustre_net.h;h=db7d8fad24de71f1c8525b2b58859d17a919e9b5;hp=da2e1cc0675dbb9a19bac3d1b43b7933e44e7aa5;hb=3998a8e474b58a5bb4bc47b620adc836c27ab70d;hpb=dad106e1272f2bae4920c081f56885efee274c57 diff --git a/lustre/include/lustre_net.h b/lustre/include/lustre_net.h index da2e1cc..db7d8fa 100644 --- a/lustre/include/lustre_net.h +++ b/lustre/include/lustre_net.h @@ -27,7 +27,7 @@ * Copyright (c) 2007, 2010, Oracle and/or its affiliates. All rights reserved. * Use is subject to license terms. * - * Copyright (c) 2010, 2012, Intel Corporation. + * Copyright (c) 2010, 2014, Intel Corporation. */ /* * This file is part of Lustre, http://www.lustre.org/ @@ -55,19 +55,9 @@ * @{ */ -#if defined(__linux__) -#include -#elif defined(__APPLE__) -#include -#elif defined(__WINNT__) -#include -#else -#error Unsupported operating system. -#endif - #include -// #include -#include +#include +#include #include #include #include @@ -75,7 +65,6 @@ #include #include #include - #include #include @@ -108,30 +97,28 @@ */ #define PTLRPC_MAX_BRW_BITS (LNET_MTU_BITS + PTLRPC_BULK_OPS_BITS) #define PTLRPC_MAX_BRW_SIZE (1 << PTLRPC_MAX_BRW_BITS) -#define PTLRPC_MAX_BRW_PAGES (PTLRPC_MAX_BRW_SIZE >> CFS_PAGE_SHIFT) +#define PTLRPC_MAX_BRW_PAGES (PTLRPC_MAX_BRW_SIZE >> PAGE_CACHE_SHIFT) #define ONE_MB_BRW_SIZE (1 << LNET_MTU_BITS) #define MD_MAX_BRW_SIZE (1 << LNET_MTU_BITS) -#define MD_MAX_BRW_PAGES (MD_MAX_BRW_SIZE >> CFS_PAGE_SHIFT) +#define MD_MAX_BRW_PAGES (MD_MAX_BRW_SIZE >> PAGE_CACHE_SHIFT) #define DT_MAX_BRW_SIZE PTLRPC_MAX_BRW_SIZE -#define DT_MAX_BRW_PAGES (DT_MAX_BRW_SIZE >> CFS_PAGE_SHIFT) +#define DT_MAX_BRW_PAGES (DT_MAX_BRW_SIZE >> PAGE_CACHE_SHIFT) #define OFD_MAX_BRW_SIZE (1 << LNET_MTU_BITS) /* When PAGE_SIZE is a constant, we can check our arithmetic here with cpp! */ -#ifdef __KERNEL__ -# if ((PTLRPC_MAX_BRW_PAGES & (PTLRPC_MAX_BRW_PAGES - 1)) != 0) -# error "PTLRPC_MAX_BRW_PAGES isn't a power of two" -# endif -# if (PTLRPC_MAX_BRW_SIZE != (PTLRPC_MAX_BRW_PAGES * CFS_PAGE_SIZE)) -# error "PTLRPC_MAX_BRW_SIZE isn't PTLRPC_MAX_BRW_PAGES * CFS_PAGE_SIZE" -# endif -# if (PTLRPC_MAX_BRW_SIZE > LNET_MTU * PTLRPC_BULK_OPS_COUNT) -# error "PTLRPC_MAX_BRW_SIZE too big" -# endif -# if (PTLRPC_MAX_BRW_PAGES > LNET_MAX_IOV * PTLRPC_BULK_OPS_COUNT) -# error "PTLRPC_MAX_BRW_PAGES too big" -# endif -#endif /* __KERNEL__ */ +#if ((PTLRPC_MAX_BRW_PAGES & (PTLRPC_MAX_BRW_PAGES - 1)) != 0) +# error "PTLRPC_MAX_BRW_PAGES isn't a power of two" +#endif +#if (PTLRPC_MAX_BRW_SIZE != (PTLRPC_MAX_BRW_PAGES * PAGE_CACHE_SIZE)) +# error "PTLRPC_MAX_BRW_SIZE isn't PTLRPC_MAX_BRW_PAGES * PAGE_CACHE_SIZE" +#endif +#if (PTLRPC_MAX_BRW_SIZE > LNET_MTU * PTLRPC_BULK_OPS_COUNT) +# error "PTLRPC_MAX_BRW_SIZE too big" +#endif +#if (PTLRPC_MAX_BRW_PAGES > LNET_MAX_IOV * PTLRPC_BULK_OPS_COUNT) +# error "PTLRPC_MAX_BRW_PAGES too big" +#endif #define PTLRPC_NTHRS_INIT 2 @@ -265,7 +252,7 @@ #define LDLM_THR_FACTOR 8 #define LDLM_NTHRS_INIT PTLRPC_NTHRS_INIT #define LDLM_NTHRS_BASE 24 -#define LDLM_NTHRS_MAX (cfs_num_online_cpus() == 1 ? 64 : 128) +#define LDLM_NTHRS_MAX (num_online_cpus() == 1 ? 64 : 128) #define LDLM_BL_THREADS LDLM_NTHRS_AUTO_INIT #define LDLM_CLIENT_NBUFS 1 @@ -380,8 +367,8 @@ * include linkea (4K maxim), together with other updates, we set it to 9K: * lustre_msg + ptlrpc_body + UPDATE_BUF_SIZE (8K) */ -#define MDS_OUT_MAXREQSIZE (9 * 1024) -#define MDS_OUT_MAXREPSIZE MDS_MAXREPSIZE +#define OUT_MAXREQSIZE (9 * 1024) +#define OUT_MAXREPSIZE MDS_MAXREPSIZE /** MDS_BUFSIZE = max_reqsize (w/o LOV EA) + max sptlrpc payload size */ #define MDS_BUFSIZE max(MDS_MAXREQSIZE + SPTLRPC_MAX_PAYLOAD, \ @@ -407,11 +394,11 @@ 160 * 1024) /** - * MDS_OUT_BUFSIZE = max_out_reqsize + max sptlrpc payload (~1K) which is + * OUT_BUFSIZE = max_out_reqsize + max sptlrpc payload (~1K) which is * about 10K, for the same reason as MDS_REG_BUFSIZE, we also give some * extra bytes to each request buffer to improve buffer utilization rate. */ -#define MDS_OUT_BUFSIZE max(MDS_OUT_MAXREQSIZE + SPTLRPC_MAX_PAYLOAD, \ +#define OUT_BUFSIZE max(OUT_MAXREQSIZE + SPTLRPC_MAX_PAYLOAD, \ 24 * 1024) /** FLD_MAXREQSIZE == lustre_msg + __u32 padding + ptlrpc_body + opc */ @@ -465,7 +452,7 @@ */ /* depress threads factor for VM with small memory size */ #define OSS_THR_FACTOR min_t(int, 8, \ - CFS_NUM_CACHEPAGES >> (28 - CFS_PAGE_SHIFT)) + NUM_CACHEPAGES >> (28 - PAGE_CACHE_SHIFT)) #define OSS_NTHRS_INIT (PTLRPC_NTHRS_INIT + 1) #define OSS_NTHRS_BASE 64 #define OSS_NTHRS_MAX 512 @@ -494,7 +481,7 @@ /** * FIEMAP request can be 4K+ for now */ -#define OST_MAXREQSIZE (5 * 1024) +#define OST_MAXREQSIZE (16 * 1024) #define OST_IO_MAXREQSIZE max_t(int, OST_MAXREQSIZE, \ (((_OST_MAXREQSIZE_SUM - 1) | (1024 - 1)) + 1)) @@ -513,20 +500,25 @@ /* Macro to hide a typecast. */ #define ptlrpc_req_async_args(req) ((void *)&req->rq_async_args) +struct ptlrpc_replay_async_args { + int praa_old_state; + int praa_old_status; +}; + /** * Structure to single define portal connection. */ struct ptlrpc_connection { - /** linkage for connections hash table */ - cfs_hlist_node_t c_hash; - /** Our own lnet nid for this connection */ - lnet_nid_t c_self; - /** Remote side nid for this connection */ - lnet_process_id_t c_peer; - /** UUID of the other side */ - struct obd_uuid c_remote_uuid; - /** reference counter for this connection */ - cfs_atomic_t c_refcount; + /** linkage for connections hash table */ + struct hlist_node c_hash; + /** Our own lnet nid for this connection */ + lnet_nid_t c_self; + /** Remote side nid for this connection */ + lnet_process_id_t c_peer; + /** UUID of the other side */ + struct obd_uuid c_remote_uuid; + /** reference counter for this connection */ + atomic_t c_refcount; }; /** Client definition for PortalRPC */ @@ -574,26 +566,26 @@ typedef int (*set_producer_func)(struct ptlrpc_request_set *, void *); * returned. */ struct ptlrpc_request_set { - cfs_atomic_t set_refcount; + atomic_t set_refcount; /** number of in queue requests */ - cfs_atomic_t set_new_count; + atomic_t set_new_count; /** number of uncompleted requests */ - cfs_atomic_t set_remaining; + atomic_t set_remaining; /** wait queue to wait on for request events */ - cfs_waitq_t set_waitq; - cfs_waitq_t *set_wakeup_ptr; + wait_queue_head_t set_waitq; + wait_queue_head_t *set_wakeup_ptr; /** List of requests in the set */ - cfs_list_t set_requests; + struct list_head set_requests; /** * List of completion callbacks to be called when the set is completed * This is only used if \a set_interpret is NULL. * Links struct ptlrpc_set_cbdata. */ - cfs_list_t set_cblist; + struct list_head set_cblist; /** Completion callback, if only one. */ - set_interpreter_func set_interpret; + set_interpreter_func set_interpret; /** opaq argument passed to completion \a set_interpret callback. */ - void *set_arg; + void *set_arg; /** * Lock for \a set_new_requests manipulations * locked so that any old caller can communicate requests to @@ -601,29 +593,29 @@ struct ptlrpc_request_set { */ spinlock_t set_new_req_lock; /** List of new yet unsent requests. Only used with ptlrpcd now. */ - cfs_list_t set_new_requests; + struct list_head set_new_requests; /** rq_status of requests that have been freed already */ - int set_rc; + int set_rc; /** Additional fields used by the flow control extension */ /** Maximum number of RPCs in flight */ - int set_max_inflight; + int set_max_inflight; /** Callback function used to generate RPCs */ - set_producer_func set_producer; + set_producer_func set_producer; /** opaq argument passed to the producer callback */ - void *set_producer_arg; + void *set_producer_arg; }; /** * Description of a single ptrlrpc_set callback */ struct ptlrpc_set_cbdata { - /** List linkage item */ - cfs_list_t psc_item; - /** Pointer to interpreting function */ - set_interpreter_func psc_interpret; - /** Opaq argument to pass to the callback */ - void *psc_data; + /** List linkage item */ + struct list_head psc_item; + /** Pointer to interpreting function */ + set_interpreter_func psc_interpret; + /** Opaq argument to pass to the callback */ + void *psc_data; }; struct ptlrpc_bulk_desc; @@ -650,20 +642,20 @@ struct ptlrpc_cb_id { * added to the state for replay/failover consistency guarantees. */ struct ptlrpc_reply_state { - /** Callback description */ - struct ptlrpc_cb_id rs_cb_id; - /** Linkage for list of all reply states in a system */ - cfs_list_t rs_list; - /** Linkage for list of all reply states on same export */ - cfs_list_t rs_exp_list; - /** Linkage for list of all reply states for same obd */ - cfs_list_t rs_obd_list; + /** Callback description */ + struct ptlrpc_cb_id rs_cb_id; + /** Linkage for list of all reply states in a system */ + struct list_head rs_list; + /** Linkage for list of all reply states on same export */ + struct list_head rs_exp_list; + /** Linkage for list of all reply states for same obd */ + struct list_head rs_obd_list; #if RS_DEBUG - cfs_list_t rs_debug_list; + struct list_head rs_debug_list; #endif - /** A spinlock to protect the reply state flags */ + /** A spinlock to protect the reply state flags */ spinlock_t rs_lock; - /** Reply state flags */ + /** Reply state flags */ unsigned long rs_difficult:1; /* ACK/commit stuff */ unsigned long rs_no_ack:1; /* no ACK, even for difficult requests */ @@ -683,16 +675,16 @@ struct ptlrpc_reply_state { __u64 rs_transno; /** xid */ __u64 rs_xid; - struct obd_export *rs_export; + struct obd_export *rs_export; struct ptlrpc_service_part *rs_svcpt; - /** Lnet metadata handle for the reply */ - lnet_handle_md_t rs_md_h; - cfs_atomic_t rs_refcount; - - /** Context for the sevice thread */ - struct ptlrpc_svc_ctx *rs_svc_ctx; - /** Reply buffer (actually sent to the client), encoded if needed */ - struct lustre_msg *rs_repbuf; /* wrapper */ + /** Lnet metadata handle for the reply */ + lnet_handle_md_t rs_md_h; + atomic_t rs_refcount; + + /** Context for the sevice thread */ + struct ptlrpc_svc_ctx *rs_svc_ctx; + /** Reply buffer (actually sent to the client), encoded if needed */ + struct lustre_msg *rs_repbuf; /* wrapper */ /** Size of the reply buffer */ int rs_repbuf_len; /* wrapper buf length */ /** Size of the reply message */ @@ -729,6 +721,9 @@ enum rq_phase { typedef int (*ptlrpc_interpterer_t)(const struct lu_env *env, struct ptlrpc_request *req, void *arg, int rc); +/** Type of request resend call-back */ +typedef void (*ptlrpc_resend_cb_t)(struct ptlrpc_request *req, + void *arg); /** * Definition of request pool structure. @@ -738,13 +733,13 @@ typedef int (*ptlrpc_interpterer_t)(const struct lu_env *env, */ struct ptlrpc_request_pool { /** Locks the list */ - spinlock_t prp_lock; - /** list of ptlrpc_request structs */ - cfs_list_t prp_req_list; - /** Maximum message size that would fit into a rquest from this pool */ - int prp_rq_size; - /** Function to allocate more requests for this pool */ - void (*prp_populate)(struct ptlrpc_request_pool *, int); + spinlock_t prp_lock; + /** list of ptlrpc_request structs */ + struct list_head prp_req_list; + /** Maximum message size that would fit into a rquest from this pool */ + int prp_rq_size; + /** Function to allocate more requests for this pool */ + void (*prp_populate)(struct ptlrpc_request_pool *, int); }; struct lu_context; @@ -787,6 +782,18 @@ enum ptlrpc_nrs_ctl { }; /** + * ORR policy operations + */ +enum nrs_ctl_orr { + NRS_CTL_ORR_RD_QUANTUM = PTLRPC_NRS_CTL_1ST_POL_SPEC, + NRS_CTL_ORR_WR_QUANTUM, + NRS_CTL_ORR_RD_OFF_TYPE, + NRS_CTL_ORR_WR_OFF_TYPE, + NRS_CTL_ORR_RD_SUPP_REQ, + NRS_CTL_ORR_WR_SUPP_REQ, +}; + +/** * NRS policy operations. * * These determine the behaviour of a policy, and are called in response to @@ -810,10 +817,12 @@ struct ptlrpc_nrs_pol_ops { * initialize their resources here; this operation is optional. * * \param[in,out] policy The policy being started + * \param[in,out] arg A generic char buffer * * \see nrs_policy_start_locked() */ - int (*op_policy_start) (struct ptlrpc_nrs_policy *policy); + int (*op_policy_start) (struct ptlrpc_nrs_policy *policy, + char *arg); /** * Called when deactivating a policy via lprocfs; policies deallocate * their resources here; this operation is optional @@ -949,7 +958,7 @@ struct ptlrpc_nrs_pol_ops { * \a nrq * \param[in,out] nrq The request * - * \pre spin_is_locked(&svcpt->scp_req_lock) + * \pre assert_spin_locked(&svcpt->scp_req_lock) * * \see ptlrpc_nrs_req_stop_nolock() */ @@ -1044,7 +1053,7 @@ struct ptlrpc_nrs { /** * List of registered policies */ - cfs_list_t nrs_policy_list; + struct list_head nrs_policy_list; /** * List of policies with queued requests. Policies that have any * outstanding requests are queued here, and this list is queried @@ -1053,7 +1062,7 @@ struct ptlrpc_nrs { * point transition away from the * ptlrpc_nrs_pol_state::NRS_POL_STATE_STARTED state are drained. */ - cfs_list_t nrs_policy_queued; + struct list_head nrs_policy_queued; /** * Service partition for this NRS head */ @@ -1091,6 +1100,10 @@ struct ptlrpc_nrs { * unregistration */ unsigned nrs_stopping:1; + /** + * NRS policy is throttling reqeust + */ + unsigned nrs_throttling:1; }; #define NRS_POL_NAME_MAX 16 @@ -1134,7 +1147,7 @@ struct ptlrpc_nrs_pol_conf { * different module to the one the NRS framework is held within * (currently ptlrpc), should set this field to THIS_MODULE. */ - cfs_module_t *nc_owner; + struct module *nc_owner; /** * Policy registration flags; a bitmast of \e nrs_policy_flags */ @@ -1155,7 +1168,7 @@ struct ptlrpc_nrs_pol_desc { /** * Link into nrs_core::nrs_policies */ - cfs_list_t pd_list; + struct list_head pd_list; /** * NRS operations for this policy */ @@ -1209,7 +1222,7 @@ struct ptlrpc_nrs_pol_desc { * then unregistration and lprocfs operations will be properly * serialized. */ - cfs_module_t *pd_owner; + struct module *pd_owner; /** * Bitmask of \e nrs_policy_flags */ @@ -1217,7 +1230,7 @@ struct ptlrpc_nrs_pol_desc { /** * # of references on this descriptor */ - cfs_atomic_t pd_refs; + atomic_t pd_refs; }; /** @@ -1291,12 +1304,12 @@ struct ptlrpc_nrs_policy { * Linkage into the NRS head's list of policies, * ptlrpc_nrs:nrs_policy_list */ - cfs_list_t pol_list; + struct list_head pol_list; /** * Linkage into the NRS head's list of policies with enqueued * requests ptlrpc_nrs:nrs_policy_queued */ - cfs_list_t pol_list_queued; + struct list_head pol_list_queued; /** * Current state of this policy */ @@ -1398,7 +1411,7 @@ struct nrs_fifo_head { /** * List of queued requests. */ - cfs_list_t fh_list; + struct list_head fh_list; /** * For debugging purposes. */ @@ -1406,13 +1419,285 @@ struct nrs_fifo_head { }; struct nrs_fifo_req { - cfs_list_t fr_list; + struct list_head fr_list; __u64 fr_sequence; }; /** @} fifo */ /** + * \name CRR-N + * + * CRR-N, Client Round Robin over NIDs + * @{ + */ + +/** + * private data structure for CRR-N NRS + */ +struct nrs_crrn_net { + struct ptlrpc_nrs_resource cn_res; + cfs_binheap_t *cn_binheap; + cfs_hash_t *cn_cli_hash; + /** + * Used when a new scheduling round commences, in order to synchronize + * all clients with the new round number. + */ + __u64 cn_round; + /** + * Determines the relevant ordering amongst request batches within a + * scheduling round. + */ + __u64 cn_sequence; + /** + * Round Robin quantum; the maximum number of RPCs that each request + * batch for each client can have in a scheduling round. + */ + __u16 cn_quantum; +}; + +/** + * Object representing a client in CRR-N, as identified by its NID + */ +struct nrs_crrn_client { + struct ptlrpc_nrs_resource cc_res; + struct hlist_node cc_hnode; + lnet_nid_t cc_nid; + /** + * The round number against which this client is currently scheduling + * requests. + */ + __u64 cc_round; + /** + * The sequence number used for requests scheduled by this client during + * the current round number. + */ + __u64 cc_sequence; + atomic_t cc_ref; + /** + * Round Robin quantum; the maximum number of RPCs the client is allowed + * to schedule in a single batch of each round. + */ + __u16 cc_quantum; + /** + * # of pending requests for this client, on all existing rounds + */ + __u16 cc_active; +}; + +/** + * CRR-N NRS request definition + */ +struct nrs_crrn_req { + /** + * Round number for this request; shared with all other requests in the + * same batch. + */ + __u64 cr_round; + /** + * Sequence number for this request; shared with all other requests in + * the same batch. + */ + __u64 cr_sequence; +}; + +/** + * CRR-N policy operations. + */ +enum nrs_ctl_crr { + /** + * Read the RR quantum size of a CRR-N policy. + */ + NRS_CTL_CRRN_RD_QUANTUM = PTLRPC_NRS_CTL_1ST_POL_SPEC, + /** + * Write the RR quantum size of a CRR-N policy. + */ + NRS_CTL_CRRN_WR_QUANTUM, +}; + +/** @} CRR-N */ + +/** + * \name ORR/TRR + * + * ORR/TRR (Object-based Round Robin/Target-based Round Robin) NRS policies + * @{ + */ + +/** + * Lower and upper byte offsets of a brw RPC + */ +struct nrs_orr_req_range { + __u64 or_start; + __u64 or_end; +}; + +/** + * RPC types supported by the ORR/TRR policies + */ +enum nrs_orr_supp { + NOS_OST_READ = (1 << 0), + NOS_OST_WRITE = (1 << 1), + NOS_OST_RW = (NOS_OST_READ | NOS_OST_WRITE), + /** + * Default value for policies. + */ + NOS_DFLT = NOS_OST_READ +}; + +/** + * As unique keys for grouping RPCs together, we use the object's OST FID for + * the ORR policy, and the OST index for the TRR policy. + * + * XXX: We waste some space for TRR policy instances by using a union, but it + * allows to consolidate some of the code between ORR and TRR, and these + * policies will probably eventually merge into one anyway. + */ +struct nrs_orr_key { + union { + /** object FID for ORR */ + struct lu_fid ok_fid; + /** OST index for TRR */ + __u32 ok_idx; + }; +}; + +/** + * The largest base string for unique hash/slab object names is + * "nrs_orr_reg_", so 13 characters. We add 3 to this to be used for the CPT + * id number, so this _should_ be more than enough for the maximum number of + * CPTs on any system. If it does happen that this statement is incorrect, + * nrs_orr_genobjname() will inevitably yield a non-unique name and cause + * kmem_cache_create() to complain (on Linux), so the erroneous situation + * will hopefully not go unnoticed. + */ +#define NRS_ORR_OBJ_NAME_MAX (sizeof("nrs_orr_reg_") + 3) + +/** + * private data structure for ORR and TRR NRS + */ +struct nrs_orr_data { + struct ptlrpc_nrs_resource od_res; + cfs_binheap_t *od_binheap; + cfs_hash_t *od_obj_hash; + struct kmem_cache *od_cache; + /** + * Used when a new scheduling round commences, in order to synchronize + * all object or OST batches with the new round number. + */ + __u64 od_round; + /** + * Determines the relevant ordering amongst request batches within a + * scheduling round. + */ + __u64 od_sequence; + /** + * RPC types that are currently supported. + */ + enum nrs_orr_supp od_supp; + /** + * Round Robin quantum; the maxium number of RPCs that each request + * batch for each object or OST can have in a scheduling round. + */ + __u16 od_quantum; + /** + * Whether to use physical disk offsets or logical file offsets. + */ + bool od_physical; + /** + * XXX: We need to provide a persistently allocated string to hold + * unique object names for this policy, since in currently supported + * versions of Linux by Lustre, kmem_cache_create() just sets a pointer + * to the name string provided. kstrdup() is used in the version of + * kmeme_cache_create() in current Linux mainline, so we may be able to + * remove this in the future. + */ + char od_objname[NRS_ORR_OBJ_NAME_MAX]; +}; + +/** + * Represents a backend-fs object or OST in the ORR and TRR policies + * respectively + */ +struct nrs_orr_object { + struct ptlrpc_nrs_resource oo_res; + struct hlist_node oo_hnode; + /** + * The round number against which requests are being scheduled for this + * object or OST + */ + __u64 oo_round; + /** + * The sequence number used for requests scheduled for this object or + * OST during the current round number. + */ + __u64 oo_sequence; + /** + * The key of the object or OST for which this structure instance is + * scheduling RPCs + */ + struct nrs_orr_key oo_key; + long oo_ref; + /** + * Round Robin quantum; the maximum number of RPCs that are allowed to + * be scheduled for the object or OST in a single batch of each round. + */ + __u16 oo_quantum; + /** + * # of pending requests for this object or OST, on all existing rounds + */ + __u16 oo_active; +}; + +/** + * ORR/TRR NRS request definition + */ +struct nrs_orr_req { + /** + * The offset range this request covers + */ + struct nrs_orr_req_range or_range; + /** + * Round number for this request; shared with all other requests in the + * same batch. + */ + __u64 or_round; + /** + * Sequence number for this request; shared with all other requests in + * the same batch. + */ + __u64 or_sequence; + /** + * For debugging purposes. + */ + struct nrs_orr_key or_key; + /** + * An ORR policy instance has filled in request information while + * enqueueing the request on the service partition's regular NRS head. + */ + unsigned int or_orr_set:1; + /** + * A TRR policy instance has filled in request information while + * enqueueing the request on the service partition's regular NRS head. + */ + unsigned int or_trr_set:1; + /** + * Request offset ranges have been filled in with logical offset + * values. + */ + unsigned int or_logical_set:1; + /** + * Request offset ranges have been filled in with physical offset + * values. + */ + unsigned int or_physical_set:1; +}; + +/** @} ORR/TRR */ + +#include + +/** * NRS request * * Instances of this object exist embedded within ptlrpc_request; the main @@ -1447,6 +1732,16 @@ struct ptlrpc_nrs_request { * Fields for the FIFO policy */ struct nrs_fifo_req fifo; + /** + * CRR-N request defintion + */ + struct nrs_crrn_req crr; + /** ORR and TRR share the same request definition */ + struct nrs_orr_req orr; + /** + * TBF request definition + */ + struct nrs_tbf_req tbf; } nr_u; /** * Externally-registering policies may want to use this to allocate @@ -1482,6 +1777,167 @@ struct ptlrpc_hpreq_ops { void (*hpreq_fini)(struct ptlrpc_request *); }; +struct ptlrpc_cli_req { + /** For bulk requests on client only: bulk descriptor */ + struct ptlrpc_bulk_desc *cr_bulk; + /** optional time limit for send attempts */ + cfs_duration_t cr_delay_limit; + /** time request was first queued */ + cfs_time_t cr_queued_time; + /** request sent timeval */ + struct timeval cr_sent_tv; + /** time for request really sent out */ + time_t cr_sent_out; + /** when req reply unlink must finish. */ + time_t cr_reply_deadline; + /** when req bulk unlink must finish. */ + time_t cr_bulk_deadline; + /** Portal to which this request would be sent */ + short cr_req_ptl; + /** Portal where to wait for reply and where reply would be sent */ + short cr_rep_ptl; + /** request resending number */ + unsigned int cr_resend_nr; + /** What was import generation when this request was sent */ + int cr_imp_gen; + enum lustre_imp_state cr_send_state; + /** Per-request waitq introduced by bug 21938 for recovery waiting */ + wait_queue_head_t cr_set_waitq; + /** Link item for request set lists */ + struct list_head cr_set_chain; + /** link to waited ctx */ + struct list_head cr_ctx_chain; + + /** client's half ctx */ + struct ptlrpc_cli_ctx *cr_cli_ctx; + /** Link back to the request set */ + struct ptlrpc_request_set *cr_set; + /** outgoing request MD handle */ + lnet_handle_md_t cr_req_md_h; + /** request-out callback parameter */ + struct ptlrpc_cb_id cr_req_cbid; + /** incoming reply MD handle */ + lnet_handle_md_t cr_reply_md_h; + wait_queue_head_t cr_reply_waitq; + /** reply callback parameter */ + struct ptlrpc_cb_id cr_reply_cbid; + /** Async completion handler, called when reply is received */ + ptlrpc_interpterer_t cr_reply_interp; + /** Resend handler, called when request is resend to update RPC data */ + ptlrpc_resend_cb_t cr_resend_cb; + /** Async completion context */ + union ptlrpc_async_args cr_async_args; + /** Opaq data for replay and commit callbacks. */ + void *cr_cb_data; + /** + * Commit callback, called when request is committed and about to be + * freed. + */ + void (*cr_commit_cb)(struct ptlrpc_request *); + /** Replay callback, called after request is replayed at recovery */ + void (*cr_replay_cb)(struct ptlrpc_request *); +}; + +/** client request member alias */ +/* NB: these alias should NOT be used by any new code, instead they should + * be removed step by step to avoid potential abuse */ +#define rq_bulk rq_cli.cr_bulk +#define rq_delay_limit rq_cli.cr_delay_limit +#define rq_queued_time rq_cli.cr_queued_time +#define rq_sent_tv rq_cli.cr_sent_tv +#define rq_real_sent rq_cli.cr_sent_out +#define rq_reply_deadline rq_cli.cr_reply_deadline +#define rq_bulk_deadline rq_cli.cr_bulk_deadline +#define rq_nr_resend rq_cli.cr_resend_nr +#define rq_request_portal rq_cli.cr_req_ptl +#define rq_reply_portal rq_cli.cr_rep_ptl +#define rq_import_generation rq_cli.cr_imp_gen +#define rq_send_state rq_cli.cr_send_state +#define rq_set_chain rq_cli.cr_set_chain +#define rq_ctx_chain rq_cli.cr_ctx_chain +#define rq_set rq_cli.cr_set +#define rq_set_waitq rq_cli.cr_set_waitq +#define rq_cli_ctx rq_cli.cr_cli_ctx +#define rq_req_md_h rq_cli.cr_req_md_h +#define rq_req_cbid rq_cli.cr_req_cbid +#define rq_reply_md_h rq_cli.cr_reply_md_h +#define rq_reply_waitq rq_cli.cr_reply_waitq +#define rq_reply_cbid rq_cli.cr_reply_cbid +#define rq_interpret_reply rq_cli.cr_reply_interp +#define rq_resend_cb rq_cli.cr_resend_cb +#define rq_async_args rq_cli.cr_async_args +#define rq_cb_data rq_cli.cr_cb_data +#define rq_commit_cb rq_cli.cr_commit_cb +#define rq_replay_cb rq_cli.cr_replay_cb + +struct ptlrpc_srv_req { + /** initial thread servicing this request */ + struct ptlrpc_thread *sr_svc_thread; + /** + * Server side list of incoming unserved requests sorted by arrival + * time. Traversed from time to time to notice about to expire + * requests and sent back "early replies" to clients to let them + * know server is alive and well, just very busy to service their + * requests in time + */ + struct list_head sr_timed_list; + /** server-side per-export list */ + struct list_head sr_exp_list; + /** server-side history, used for debuging purposes. */ + struct list_head sr_hist_list; + /** history sequence # */ + __u64 sr_hist_seq; + /** the index of service's srv_at_array into which request is linked */ + time_t sr_at_index; + /** authed uid */ + uid_t sr_auth_uid; + /** authed uid mapped to */ + uid_t sr_auth_mapped_uid; + /** RPC is generated from what part of Lustre */ + enum lustre_sec_part sr_sp_from; + /** request session context */ + struct lu_context sr_ses; + /** \addtogroup nrs + * @{ + */ + /** stub for NRS request */ + struct ptlrpc_nrs_request sr_nrq; + /** @} nrs */ + /** request arrival time */ + struct timeval sr_arrival_time; + /** server's half ctx */ + struct ptlrpc_svc_ctx *sr_svc_ctx; + /** (server side), pointed directly into req buffer */ + struct ptlrpc_user_desc *sr_user_desc; + /** separated reply state */ + struct ptlrpc_reply_state *sr_reply_state; + /** server-side hp handlers */ + struct ptlrpc_hpreq_ops *sr_ops; + /** incoming request buffer */ + struct ptlrpc_request_buffer_desc *sr_rqbd; +}; + +/** server request member alias */ +/* NB: these alias should NOT be used by any new code, instead they should + * be removed step by step to avoid potential abuse */ +#define rq_svc_thread rq_srv.sr_svc_thread +#define rq_timed_list rq_srv.sr_timed_list +#define rq_exp_list rq_srv.sr_exp_list +#define rq_history_list rq_srv.sr_hist_list +#define rq_history_seq rq_srv.sr_hist_seq +#define rq_at_index rq_srv.sr_at_index +#define rq_auth_uid rq_srv.sr_auth_uid +#define rq_auth_mapped_uid rq_srv.sr_auth_mapped_uid +#define rq_sp_from rq_srv.sr_sp_from +#define rq_session rq_srv.sr_ses +#define rq_nrq rq_srv.sr_nrq +#define rq_arrival_time rq_srv.sr_arrival_time +#define rq_reply_state rq_srv.sr_reply_state +#define rq_svc_ctx rq_srv.sr_svc_ctx +#define rq_user_desc rq_srv.sr_user_desc +#define rq_ops rq_srv.sr_ops +#define rq_rqbd rq_srv.sr_rqbd + /** * Represents remote procedure call. * @@ -1490,47 +1946,19 @@ struct ptlrpc_hpreq_ops { */ struct ptlrpc_request { /* Request type: one of PTL_RPC_MSG_* */ - int rq_type; + int rq_type; /** Result of request processing */ - int rq_status; - /** - * Linkage item through which this request is included into - * sending/delayed lists on client and into rqbd list on server - */ - cfs_list_t rq_list; - /** - * Server side list of incoming unserved requests sorted by arrival - * time. Traversed from time to time to notice about to expire - * requests and sent back "early replies" to clients to let them - * know server is alive and well, just very busy to service their - * requests in time - */ - cfs_list_t rq_timed_list; - /** server-side history, used for debuging purposes. */ - cfs_list_t rq_history_list; - /** server-side per-export list */ - cfs_list_t rq_exp_list; - /** server-side hp handlers */ - struct ptlrpc_hpreq_ops *rq_ops; - - /** initial thread servicing this request */ - struct ptlrpc_thread *rq_svc_thread; - - /** history sequence # */ - __u64 rq_history_seq; - /** \addtogroup nrs - * @{ + int rq_status; + /** + * Linkage item through which this request is included into + * sending/delayed lists on client and into rqbd list on server */ - /** stub for NRS request */ - struct ptlrpc_nrs_request rq_nrq; - /** @} nrs */ - /** the index of service's srv_at_array into which request is linked */ - time_t rq_at_index; - /** Lock to protect request flags and some other important bits, like - * rq_list - */ - spinlock_t rq_lock; - /** client-side flags are serialized by rq_lock */ + struct list_head rq_list; + /** Lock to protect request flags and some other important bits, like + * rq_list + */ + spinlock_t rq_lock; + /** client-side flags are serialized by rq_lock @{ */ unsigned int rq_intr:1, rq_replied:1, rq_err:1, rq_timedout:1, rq_resend:1, rq_restart:1, /** @@ -1545,35 +1973,40 @@ struct ptlrpc_request { rq_replay:1, rq_no_resend:1, rq_waiting:1, rq_receiving_reply:1, rq_no_delay:1, rq_net_err:1, rq_wait_ctx:1, - rq_early:1, rq_must_unlink:1, - rq_memalloc:1, /* req originated from "kswapd" */ - /* server-side flags */ - rq_packed_final:1, /* packed final reply */ - rq_hp:1, /* high priority RPC */ - rq_at_linked:1, /* link into service's srv_at_array */ - rq_reply_truncate:1, - rq_committed:1, - /* whether the "rq_set" is a valid one */ - rq_invalid_rqset:1, + rq_early:1, + rq_req_unlinked:1, /* unlinked request buffer from lnet */ + rq_reply_unlinked:1, /* unlinked reply buffer from lnet */ + rq_memalloc:1, /* req originated from "kswapd" */ + rq_committed:1, + rq_reply_truncated:1, + /** whether the "rq_set" is a valid one */ + rq_invalid_rqset:1, rq_generation_set:1, - /* do not resend request on -EINPROGRESS */ + /** do not resend request on -EINPROGRESS */ rq_no_retry_einprogress:1, /* allow the req to be sent if the import is in recovery * status */ - rq_allow_replay:1; - - unsigned int rq_nr_resend; - - enum rq_phase rq_phase; /* one of RQ_PHASE_* */ - enum rq_phase rq_next_phase; /* one of RQ_PHASE_* to be used next */ - cfs_atomic_t rq_refcount;/* client-side refcount for SENT race, - server-side refcounf for multiple replies */ + rq_allow_replay:1, + /* bulk request, sent to server, but uncommitted */ + rq_unstable:1; + /** @} */ - /** Portal to which this request would be sent */ - short rq_request_portal; /* XXX FIXME bug 249 */ - /** Portal where to wait for reply and where reply would be sent */ - short rq_reply_portal; /* XXX FIXME bug 249 */ + /** server-side flags @{ */ + unsigned int + rq_hp:1, /**< high priority RPC */ + rq_at_linked:1, /**< link into service's srv_at_array */ + rq_packed_final:1; /**< packed final reply */ + /** @} */ + /** one of RQ_PHASE_* */ + enum rq_phase rq_phase; + /** one of RQ_PHASE_* to be used next */ + enum rq_phase rq_next_phase; + /** + * client-side refcount for SENT race, server-side refcounf + * for multiple replies + */ + atomic_t rq_refcount; /** * client-side: * !rq_truncate : # reply bytes actually received, @@ -1584,6 +2017,8 @@ struct ptlrpc_request { int rq_reqlen; /** Reply length */ int rq_replen; + /** Pool if request is from preallocated list */ + struct ptlrpc_request_pool *rq_pool; /** Request message - what client sent */ struct lustre_msg *rq_reqmsg; /** Reply message - server response */ @@ -1592,22 +2027,23 @@ struct ptlrpc_request { __u64 rq_transno; /** xid */ __u64 rq_xid; - /** - * List item to for replay list. Not yet commited requests get linked - * there. - * Also see \a rq_replay comment above. - */ - cfs_list_t rq_replay_list; - - /** - * security and encryption data - * @{ */ - struct ptlrpc_cli_ctx *rq_cli_ctx; /**< client's half ctx */ - struct ptlrpc_svc_ctx *rq_svc_ctx; /**< server's half ctx */ - cfs_list_t rq_ctx_chain; /**< link to waited ctx */ - - struct sptlrpc_flavor rq_flvr; /**< for client & server */ - enum lustre_sec_part rq_sp_from; + /** + * List item to for replay list. Not yet commited requests get linked + * there. + * Also see \a rq_replay comment above. + * It's also link chain on obd_export::exp_req_replay_queue + */ + struct list_head rq_replay_list; + /** non-shared members for client & server request*/ + union { + struct ptlrpc_cli_req rq_cli; + struct ptlrpc_srv_req rq_srv; + }; + /** + * security and encryption data + * @{ */ + /** description of flavors for client & server */ + struct sptlrpc_flavor rq_flvr; /* client/server security flags */ unsigned int @@ -1626,19 +2062,16 @@ struct ptlrpc_request { rq_pack_bulk:1, /* doesn't expect reply FIXME */ rq_no_reply:1, - rq_pill_init:1; /* pill initialized */ + rq_pill_init:1, /* pill initialized */ + rq_srv_req:1; /* server request */ - uid_t rq_auth_uid; /* authed uid */ - uid_t rq_auth_mapped_uid; /* authed uid mapped to */ - /* (server side), pointed directly into req buffer */ - struct ptlrpc_user_desc *rq_user_desc; - - /* various buffer pointers */ - struct lustre_msg *rq_reqbuf; /* req wrapper */ - char *rq_repbuf; /* rep buffer */ - struct lustre_msg *rq_repdata; /* rep wrapper msg */ - struct lustre_msg *rq_clrbuf; /* only in priv mode */ + /** various buffer pointers */ + struct lustre_msg *rq_reqbuf; /**< req wrapper */ + char *rq_repbuf; /**< rep buffer */ + struct lustre_msg *rq_repdata; /**< rep wrapper msg */ + /** only in priv mode */ + struct lustre_msg *rq_clrbuf; int rq_reqbuf_len; /* req wrapper buf len */ int rq_reqdata_len; /* req wrapper msg len */ int rq_repbuf_len; /* rep buffer len */ @@ -1647,106 +2080,37 @@ struct ptlrpc_request { int rq_clrdata_len; /* only in priv mode */ /** early replies go to offset 0, regular replies go after that */ - unsigned int rq_reply_off; - - /** @} */ - - /** Fields that help to see if request and reply were swabbed or not */ - __u32 rq_req_swab_mask; - __u32 rq_rep_swab_mask; - - /** What was import generation when this request was sent */ - int rq_import_generation; - enum lustre_imp_state rq_send_state; - - /** how many early replies (for stats) */ - int rq_early_count; - - /** client+server request */ - lnet_handle_md_t rq_req_md_h; - struct ptlrpc_cb_id rq_req_cbid; - /** optional time limit for send attempts */ - cfs_duration_t rq_delay_limit; - /** time request was first queued */ - cfs_time_t rq_queued_time; - - /* server-side... */ - /** request arrival time */ - struct timeval rq_arrival_time; - /** separated reply state */ - struct ptlrpc_reply_state *rq_reply_state; - /** incoming request buffer */ - struct ptlrpc_request_buffer_desc *rq_rqbd; - - /** client-only incoming reply */ - lnet_handle_md_t rq_reply_md_h; - cfs_waitq_t rq_reply_waitq; - struct ptlrpc_cb_id rq_reply_cbid; - - /** our LNet NID */ - lnet_nid_t rq_self; - /** Peer description (the other side) */ - lnet_process_id_t rq_peer; - /** Server-side, export on which request was received */ - struct obd_export *rq_export; - /** Client side, import where request is being sent */ - struct obd_import *rq_import; - - /** Replay callback, called after request is replayed at recovery */ - void (*rq_replay_cb)(struct ptlrpc_request *); - /** - * Commit callback, called when request is committed and about to be - * freed. - */ - void (*rq_commit_cb)(struct ptlrpc_request *); - /** Opaq data for replay and commit callbacks. */ - void *rq_cb_data; - - /** For bulk requests on client only: bulk descriptor */ - struct ptlrpc_bulk_desc *rq_bulk; + unsigned int rq_reply_off; - /** client outgoing req */ - /** - * when request/reply sent (secs), or time when request should be sent - */ - time_t rq_sent; - /** time for request really sent out */ - time_t rq_real_sent; - - /** when request must finish. volatile - * so that servers' early reply updates to the deadline aren't - * kept in per-cpu cache */ - volatile time_t rq_deadline; - /** when req reply unlink must finish. */ - time_t rq_reply_deadline; - /** when req bulk unlink must finish. */ - time_t rq_bulk_deadline; - /** - * service time estimate (secs) - * If the requestsis not served by this time, it is marked as timed out. - */ - int rq_timeout; + /** @} */ - /** Multi-rpc bits */ - /** Per-request waitq introduced by bug 21938 for recovery waiting */ - cfs_waitq_t rq_set_waitq; - /** Link item for request set lists */ - cfs_list_t rq_set_chain; - /** Link back to the request set */ - struct ptlrpc_request_set *rq_set; - /** Async completion handler, called when reply is received */ - ptlrpc_interpterer_t rq_interpret_reply; - /** Async completion context */ - union ptlrpc_async_args rq_async_args; - - /** Pool if request is from preallocated list */ - struct ptlrpc_request_pool *rq_pool; - - struct lu_context rq_session; - struct lu_context rq_recov_session; - - /** request format description */ - struct req_capsule rq_pill; + /** Fields that help to see if request and reply were swabbed or not */ + __u32 rq_req_swab_mask; + __u32 rq_rep_swab_mask; + + /** how many early replies (for stats) */ + int rq_early_count; + /** Server-side, export on which request was received */ + struct obd_export *rq_export; + /** import where request is being sent */ + struct obd_import *rq_import; + /** our LNet NID */ + lnet_nid_t rq_self; + /** Peer description (the other side) */ + lnet_process_id_t rq_peer; + /** + * service time estimate (secs) + * If the request is not served by this time, it is marked as timed out. + */ + int rq_timeout; + /** + * when request/reply sent (secs), or time when request should be sent + */ + time_t rq_sent; + /** when request must finish. */ + time_t rq_deadline; + /** request format description */ + struct req_capsule rq_pill; }; /** @@ -1797,7 +2161,7 @@ static inline bool ptlrpc_nrs_req_can_move(struct ptlrpc_request *req) /** * Returns 1 if request buffer at offset \a index was already swabbed */ -static inline int lustre_req_swabbed(struct ptlrpc_request *req, int index) +static inline int lustre_req_swabbed(struct ptlrpc_request *req, size_t index) { LASSERT(index < sizeof(req->rq_req_swab_mask) * 8); return req->rq_req_swab_mask & (1 << index); @@ -1806,7 +2170,7 @@ static inline int lustre_req_swabbed(struct ptlrpc_request *req, int index) /** * Returns 1 if request reply buffer at offset \a index was already swabbed */ -static inline int lustre_rep_swabbed(struct ptlrpc_request *req, int index) +static inline int lustre_rep_swabbed(struct ptlrpc_request *req, size_t index) { LASSERT(index < sizeof(req->rq_rep_swab_mask) * 8); return req->rq_rep_swab_mask & (1 << index); @@ -1831,7 +2195,8 @@ static inline int ptlrpc_rep_need_swab(struct ptlrpc_request *req) /** * Mark request buffer at offset \a index that it was already swabbed */ -static inline void lustre_set_req_swabbed(struct ptlrpc_request *req, int index) +static inline void lustre_set_req_swabbed(struct ptlrpc_request *req, + size_t index) { LASSERT(index < sizeof(req->rq_req_swab_mask) * 8); LASSERT((req->rq_req_swab_mask & (1 << index)) == 0); @@ -1841,7 +2206,8 @@ static inline void lustre_set_req_swabbed(struct ptlrpc_request *req, int index) /** * Mark request reply buffer at offset \a index that it was already swabbed */ -static inline void lustre_set_rep_swabbed(struct ptlrpc_request *req, int index) +static inline void lustre_set_rep_swabbed(struct ptlrpc_request *req, + size_t index) { LASSERT(index < sizeof(req->rq_rep_swab_mask) * 8); LASSERT((req->rq_rep_swab_mask & (1 << index)) == 0); @@ -1885,7 +2251,7 @@ ptlrpc_rqphase2str(struct ptlrpc_request *req) /** * Debugging functions and helpers to print request structure into debug log * @{ - */ + */ /* Spare the preprocessor, spoil the bugs. */ #define FLAG(field, str) (field ? str : "") @@ -1942,16 +2308,16 @@ do { \ * Structure that defines a single page of a bulk transfer */ struct ptlrpc_bulk_page { - /** Linkage to list of pages in a bulk */ - cfs_list_t bp_link; - /** - * Number of bytes in a page to transfer starting from \a bp_pageoffset - */ - int bp_buflen; - /** offset within a page */ - int bp_pageoffset; - /** The page itself */ - struct page *bp_page; + /** Linkage to list of pages in a bulk */ + struct list_head bp_link; + /** + * Number of bytes in a page to transfer starting from \a bp_pageoffset + */ + int bp_buflen; + /** offset within a page */ + int bp_pageoffset; + /** The page itself */ + struct page *bp_page; }; #define BULK_GET_SOURCE 0 @@ -1988,7 +2354,7 @@ struct ptlrpc_bulk_desc { struct obd_import *bd_import; /** Back pointer to the request */ struct ptlrpc_request *bd_req; - cfs_waitq_t bd_waitq; /* server side only WQ */ + wait_queue_head_t bd_waitq; /* server side only WQ */ int bd_iov_count; /* # entries in bd_iov */ int bd_max_iov; /* allocated size of bd_iov */ int bd_nob; /* # bytes covered */ @@ -2003,16 +2369,12 @@ struct ptlrpc_bulk_desc { /** array of associated MDs */ lnet_handle_md_t bd_mds[PTLRPC_BULK_OPS_COUNT]; -#if defined(__KERNEL__) /* * encrypt iov, size is either 0 or bd_iov_count. */ lnet_kiov_t *bd_enc_iov; lnet_kiov_t bd_iov[0]; -#else - lnet_md_iovec_t bd_iov[0]; -#endif }; enum { @@ -2032,7 +2394,7 @@ struct ptlrpc_thread { /** * List of active threads in svc->srv_threads */ - cfs_list_t t_link; + struct list_head t_link; /** * thread-private data (preallocated memory) */ @@ -2045,7 +2407,7 @@ struct ptlrpc_thread { /** * service thread pid */ - pid_t t_pid; + pid_t t_pid; /** * put watchdog in the structure per thread b=14840 */ @@ -2054,7 +2416,7 @@ struct ptlrpc_thread { * the svc this thread belonged to b=18582 */ struct ptlrpc_service_part *t_svcpt; - cfs_waitq_t t_ctl_waitq; + wait_queue_head_t t_ctl_waitq; struct lu_env *t_env; char t_name[PTLRPC_THR_NAME_LEN]; }; @@ -2127,23 +2489,23 @@ static inline int thread_test_and_clear_flags(struct ptlrpc_thread *thread, * More than one request can fit into the buffer. */ struct ptlrpc_request_buffer_desc { - /** Link item for rqbds on a service */ - cfs_list_t rqbd_list; - /** History of requests for this buffer */ - cfs_list_t rqbd_reqs; - /** Back pointer to service for which this buffer is registered */ - struct ptlrpc_service_part *rqbd_svcpt; - /** LNet descriptor */ - lnet_handle_md_t rqbd_md_h; - int rqbd_refcount; - /** The buffer itself */ - char *rqbd_buffer; - struct ptlrpc_cb_id rqbd_cbid; - /** - * This "embedded" request structure is only used for the - * last request to fit into the buffer - */ - struct ptlrpc_request rqbd_req; + /** Link item for rqbds on a service */ + struct list_head rqbd_list; + /** History of requests for this buffer */ + struct list_head rqbd_reqs; + /** Back pointer to service for which this buffer is registered */ + struct ptlrpc_service_part *rqbd_svcpt; + /** LNet descriptor */ + lnet_handle_md_t rqbd_md_h; + int rqbd_refcount; + /** The buffer itself */ + char *rqbd_buffer; + struct ptlrpc_cb_id rqbd_cbid; + /** + * This "embedded" request structure is only used for the + * last request to fit into the buffer + */ + struct ptlrpc_request rqbd_req; }; typedef int (*svc_handler_t)(struct ptlrpc_request *req); @@ -2194,9 +2556,9 @@ struct ptlrpc_service_ops { struct ptlrpc_service { /** serialize /proc operations */ spinlock_t srv_lock; - /** most often accessed fields */ - /** chain thru all services */ - cfs_list_t srv_list; + /** most often accessed fields */ + /** chain thru all services */ + struct list_head srv_list; /** service operations table */ struct ptlrpc_service_ops srv_ops; /** only statically allocated strings here; we don't clean them */ @@ -2204,13 +2566,13 @@ struct ptlrpc_service { /** only statically allocated strings here; we don't clean them */ char *srv_thread_name; /** service thread list */ - cfs_list_t srv_threads; + struct list_head srv_threads; /** threads # should be created for each partition on initializing */ int srv_nthrs_cpt_init; /** limit of threads number for each partition */ int srv_nthrs_cpt_limit; /** Root of /proc dir tree for this service */ - cfs_proc_dir_entry_t *srv_procroot; + struct proc_dir_entry *srv_procroot; /** Pointer to statistic data for this service */ struct lprocfs_stats *srv_stats; /** # hp per lp reqs to handle */ @@ -2285,7 +2647,7 @@ struct ptlrpc_service_part { /** # running threads */ int scp_nthrs_running; /** service threads list */ - cfs_list_t scp_threads; + struct list_head scp_threads; /** * serialize the following fields, used for protecting @@ -2302,23 +2664,23 @@ struct ptlrpc_service_part { /** # incoming reqs */ int scp_nreqs_incoming; /** request buffers to be reposted */ - cfs_list_t scp_rqbd_idle; + struct list_head scp_rqbd_idle; /** req buffers receiving */ - cfs_list_t scp_rqbd_posted; + struct list_head scp_rqbd_posted; /** incoming reqs */ - cfs_list_t scp_req_incoming; + struct list_head scp_req_incoming; /** timeout before re-posting reqs, in tick */ cfs_duration_t scp_rqbd_timeout; /** * all threads sleep on this. This wait-queue is signalled when new * incoming request arrives and when difficult reply has to be handled. */ - cfs_waitq_t scp_waitq; + wait_queue_head_t scp_waitq; /** request history */ - cfs_list_t scp_hist_reqs; + struct list_head scp_hist_reqs; /** request buffer history */ - cfs_list_t scp_hist_rqbds; + struct list_head scp_hist_rqbds; /** # request buffers in history */ int scp_hist_nrqbds; /** sequence number for request */ @@ -2357,7 +2719,7 @@ struct ptlrpc_service_part { /** reqs waiting for replies */ struct ptlrpc_at_array scp_at_array; /** early reply timer */ - cfs_timer_t scp_at_timer; + struct timer_list scp_at_timer; /** debug */ cfs_time_t scp_at_checktime; /** check early replies */ @@ -2370,17 +2732,13 @@ struct ptlrpc_service_part { */ spinlock_t scp_rep_lock __cfs_cacheline_aligned; /** all the active replies */ - cfs_list_t scp_rep_active; -#ifndef __KERNEL__ - /** replies waiting for service */ - cfs_list_t scp_rep_queue; -#endif + struct list_head scp_rep_active; /** List of free reply_states */ - cfs_list_t scp_rep_idle; + struct list_head scp_rep_idle; /** waitq to run, when adding stuff to srv_free_rs_list */ - cfs_waitq_t scp_rep_waitq; + wait_queue_head_t scp_rep_waitq; /** # 'difficult' replies */ - cfs_atomic_t scp_nreps_difficult; + atomic_t scp_nreps_difficult; }; #define ptlrpc_service_for_each_part(part, i, svc) \ @@ -2414,7 +2772,7 @@ struct ptlrpcd_ctl { */ struct ptlrpc_request_set *pc_set; /** - * Thread name used in cfs_daemonize() + * Thread name used in kthread_run() */ char pc_name[16]; /** @@ -2437,25 +2795,6 @@ struct ptlrpcd_ctl { * Record the partner index to be processed next. */ int pc_cursor; -#ifndef __KERNEL__ - /** - * Async rpcs flag to make sure that ptlrpcd_check() is called only - * once. - */ - int pc_recurred; - /** - * Currently not used. - */ - void *pc_callback; - /** - * User-space async rpcs callback. - */ - void *pc_wait_callback; - /** - * User-space check idle rpcs callback. - */ - void *pc_idle_callback; -#endif }; /* Bits for pc_flags */ @@ -2609,7 +2948,6 @@ int ptlrpc_send_reply(struct ptlrpc_request *req, int flags); int ptlrpc_reply(struct ptlrpc_request *req); int ptlrpc_send_error(struct ptlrpc_request *req, int difficult); int ptlrpc_error(struct ptlrpc_request *req); -void ptlrpc_resend_req(struct ptlrpc_request *request); int ptlrpc_at_get_net_latency(struct ptlrpc_request *req); int ptl_send_rpc(struct ptlrpc_request *request, int noreply); int ptlrpc_register_rqbd(struct ptlrpc_request_buffer_desc *rqbd); @@ -2621,6 +2959,8 @@ int ptlrpc_register_rqbd(struct ptlrpc_request_buffer_desc *rqbd); * request queues, request management, etc. * @{ */ +void ptlrpc_request_committed(struct ptlrpc_request *req, int force); + void ptlrpc_init_client(int req_portal, int rep_portal, char *name, struct ptlrpc_client *); void ptlrpc_cleanup_client(struct obd_import *imp); @@ -2628,7 +2968,6 @@ struct ptlrpc_connection *ptlrpc_uuid_to_connection(struct obd_uuid *uuid); int ptlrpc_queue_wait(struct ptlrpc_request *req); int ptlrpc_replay_req(struct ptlrpc_request *req); -int ptlrpc_unregister_reply(struct ptlrpc_request *req, int async); void ptlrpc_restart_req(struct ptlrpc_request *req); void ptlrpc_abort_inflight(struct obd_import *imp); void ptlrpc_cleanup_imp(struct obd_import *imp); @@ -2639,16 +2978,11 @@ struct ptlrpc_request_set *ptlrpc_prep_fcset(int max, set_producer_func func, void *arg); int ptlrpc_set_add_cb(struct ptlrpc_request_set *set, set_interpreter_func fn, void *data); -int ptlrpc_set_next_timeout(struct ptlrpc_request_set *); int ptlrpc_check_set(const struct lu_env *env, struct ptlrpc_request_set *set); int ptlrpc_set_wait(struct ptlrpc_request_set *); -int ptlrpc_expired_set(void *data); -void ptlrpc_interrupted_set(void *data); void ptlrpc_mark_interrupted(struct ptlrpc_request *req); void ptlrpc_set_destroy(struct ptlrpc_request_set *); void ptlrpc_set_add_req(struct ptlrpc_request_set *, struct ptlrpc_request *); -void ptlrpc_set_add_new_req(struct ptlrpcd_ctl *pc, - struct ptlrpc_request *req); void ptlrpc_free_rq_pool(struct ptlrpc_request_pool *pool); void ptlrpc_add_rqs_to_pool(struct ptlrpc_request_pool *pool, int num_rq); @@ -2695,16 +3029,16 @@ static inline void ptlrpc_free_bulk_nopin(struct ptlrpc_bulk_desc *bulk) __ptlrpc_free_bulk(bulk, 0); } void __ptlrpc_prep_bulk_page(struct ptlrpc_bulk_desc *desc, - cfs_page_t *page, int pageoffset, int len, int); + struct page *page, int pageoffset, int len, int); static inline void ptlrpc_prep_bulk_page_pin(struct ptlrpc_bulk_desc *desc, - cfs_page_t *page, int pageoffset, + struct page *page, int pageoffset, int len) { __ptlrpc_prep_bulk_page(desc, page, pageoffset, len, 1); } static inline void ptlrpc_prep_bulk_page_nopin(struct ptlrpc_bulk_desc *desc, - cfs_page_t *page, int pageoffset, + struct page *page, int pageoffset, int len) { __ptlrpc_prep_bulk_page(desc, page, pageoffset, len, 0); @@ -2808,14 +3142,12 @@ int liblustre_check_services(void *arg); void ptlrpc_daemonize(char *name); int ptlrpc_service_health_check(struct ptlrpc_service *); void ptlrpc_server_drop_request(struct ptlrpc_request *req); +void ptlrpc_request_change_export(struct ptlrpc_request *req, + struct obd_export *export); +void ptlrpc_update_export_timer(struct obd_export *exp, long extra_delay); -#ifdef __KERNEL__ int ptlrpc_hr_init(void); void ptlrpc_hr_fini(void); -#else -# define ptlrpc_hr_init() (0) -# define ptlrpc_hr_fini() do {} while(0) -#endif /** @} */ @@ -2836,14 +3168,14 @@ int ptlrpc_reconnect_import(struct obd_import *imp); /** @} */ /** - * ptlrpc msg buffer and swab interface + * ptlrpc msg buffer and swab interface * * @{ */ int ptlrpc_buf_need_swab(struct ptlrpc_request *req, const int inout, - int index); + __u32 index); void ptlrpc_buf_set_swabbed(struct ptlrpc_request *req, const int inout, - int index); + __u32 index); int ptlrpc_unpack_rep_msg(struct ptlrpc_request *req, int len); int ptlrpc_unpack_req_msg(struct ptlrpc_request *req, int len); @@ -2863,32 +3195,33 @@ int lustre_shrink_msg(struct lustre_msg *msg, int segment, unsigned int newlen, int move_data); void lustre_free_reply_state(struct ptlrpc_reply_state *rs); int __lustre_unpack_msg(struct lustre_msg *m, int len); -int lustre_msg_hdr_size(__u32 magic, int count); -int lustre_msg_size(__u32 magic, int count, __u32 *lengths); -int lustre_msg_size_v2(int count, __u32 *lengths); -int lustre_packed_msg_size(struct lustre_msg *msg); -int lustre_msg_early_size(void); -void *lustre_msg_buf_v2(struct lustre_msg_v2 *m, int n, int min_size); -void *lustre_msg_buf(struct lustre_msg *m, int n, int minlen); -int lustre_msg_buflen(struct lustre_msg *m, int n); -void lustre_msg_set_buflen(struct lustre_msg *m, int n, int len); -int lustre_msg_bufcount(struct lustre_msg *m); -char *lustre_msg_string(struct lustre_msg *m, int n, int max_len); +__u32 lustre_msg_hdr_size(__u32 magic, __u32 count); +__u32 lustre_msg_size(__u32 magic, int count, __u32 *lengths); +__u32 lustre_msg_size_v2(int count, __u32 *lengths); +__u32 lustre_packed_msg_size(struct lustre_msg *msg); +__u32 lustre_msg_early_size(void); +void *lustre_msg_buf_v2(struct lustre_msg_v2 *m, __u32 n, __u32 min_size); +void *lustre_msg_buf(struct lustre_msg *m, __u32 n, __u32 minlen); +__u32 lustre_msg_buflen(struct lustre_msg *m, __u32 n); +void lustre_msg_set_buflen(struct lustre_msg *m, __u32 n, __u32 len); +__u32 lustre_msg_bufcount(struct lustre_msg *m); +char *lustre_msg_string(struct lustre_msg *m, __u32 n, __u32 max_len); __u32 lustre_msghdr_get_flags(struct lustre_msg *msg); void lustre_msghdr_set_flags(struct lustre_msg *msg, __u32 flags); __u32 lustre_msg_get_flags(struct lustre_msg *msg); -void lustre_msg_add_flags(struct lustre_msg *msg, int flags); -void lustre_msg_set_flags(struct lustre_msg *msg, int flags); -void lustre_msg_clear_flags(struct lustre_msg *msg, int flags); +void lustre_msg_add_flags(struct lustre_msg *msg, __u32 flags); +void lustre_msg_set_flags(struct lustre_msg *msg, __u32 flags); +void lustre_msg_clear_flags(struct lustre_msg *msg, __u32 flags); __u32 lustre_msg_get_op_flags(struct lustre_msg *msg); -void lustre_msg_add_op_flags(struct lustre_msg *msg, int flags); -void lustre_msg_set_op_flags(struct lustre_msg *msg, int flags); +void lustre_msg_add_op_flags(struct lustre_msg *msg, __u32 flags); +void lustre_msg_set_op_flags(struct lustre_msg *msg, __u32 flags); struct lustre_handle *lustre_msg_get_handle(struct lustre_msg *msg); __u32 lustre_msg_get_type(struct lustre_msg *msg); __u32 lustre_msg_get_version(struct lustre_msg *msg); -void lustre_msg_add_version(struct lustre_msg *msg, int version); +void lustre_msg_add_version(struct lustre_msg *msg, __u32 version); __u32 lustre_msg_get_opc(struct lustre_msg *msg); __u64 lustre_msg_get_last_xid(struct lustre_msg *msg); +__u16 lustre_msg_get_tag(struct lustre_msg *msg); __u64 lustre_msg_get_last_committed(struct lustre_msg *msg); __u64 *lustre_msg_get_versions(struct lustre_msg *msg); __u64 lustre_msg_get_transno(struct lustre_msg *msg); @@ -2904,16 +3237,16 @@ __u32 lustre_msg_get_timeout(struct lustre_msg *msg); __u32 lustre_msg_get_service_time(struct lustre_msg *msg); char *lustre_msg_get_jobid(struct lustre_msg *msg); __u32 lustre_msg_get_cksum(struct lustre_msg *msg); -#if LUSTRE_VERSION_CODE < OBD_OCD_VERSION(2, 7, 50, 0) +#if LUSTRE_VERSION_CODE < OBD_OCD_VERSION(2, 7, 53, 0) __u32 lustre_msg_calc_cksum(struct lustre_msg *msg, int compat18); #else -# warning "remove checksum compatibility support for b1_8" __u32 lustre_msg_calc_cksum(struct lustre_msg *msg); #endif void lustre_msg_set_handle(struct lustre_msg *msg,struct lustre_handle *handle); void lustre_msg_set_type(struct lustre_msg *msg, __u32 type); void lustre_msg_set_opc(struct lustre_msg *msg, __u32 opc); void lustre_msg_set_last_xid(struct lustre_msg *msg, __u64 last_xid); +void lustre_msg_set_tag(struct lustre_msg *msg, __u16 tag); void lustre_msg_set_last_committed(struct lustre_msg *msg,__u64 last_committed); void lustre_msg_set_versions(struct lustre_msg *msg, __u64 *versions); void lustre_msg_set_transno(struct lustre_msg *msg, __u64 transno); @@ -2935,34 +3268,66 @@ lustre_shrink_reply(struct ptlrpc_request *req, int segment, req->rq_replen = lustre_shrink_msg(req->rq_repmsg, segment, newlen, move_data); } + +#ifdef LUSTRE_TRANSLATE_ERRNOS + +static inline int ptlrpc_status_hton(int h) +{ + /* + * Positive errnos must be network errnos, such as LUSTRE_EDEADLK, + * ELDLM_LOCK_ABORTED, etc. + */ + if (h < 0) + return -lustre_errno_hton(-h); + else + return h; +} + +static inline int ptlrpc_status_ntoh(int n) +{ + /* + * See the comment in ptlrpc_status_hton(). + */ + if (n < 0) + return -lustre_errno_ntoh(-n); + else + return n; +} + +#else + +#define ptlrpc_status_hton(h) (h) +#define ptlrpc_status_ntoh(n) (n) + +#endif /** @} */ /** Change request phase of \a req to \a new_phase */ static inline void ptlrpc_rqphase_move(struct ptlrpc_request *req, enum rq_phase new_phase) { - if (req->rq_phase == new_phase) - return; + if (req->rq_phase == new_phase) + return; - if (new_phase == RQ_PHASE_UNREGISTERING) { - req->rq_next_phase = req->rq_phase; - if (req->rq_import) - cfs_atomic_inc(&req->rq_import->imp_unregistering); - } + if (new_phase == RQ_PHASE_UNREGISTERING) { + req->rq_next_phase = req->rq_phase; + if (req->rq_import) + atomic_inc(&req->rq_import->imp_unregistering); + } - if (req->rq_phase == RQ_PHASE_UNREGISTERING) { - if (req->rq_import) - cfs_atomic_dec(&req->rq_import->imp_unregistering); - } + if (req->rq_phase == RQ_PHASE_UNREGISTERING) { + if (req->rq_import) + atomic_dec(&req->rq_import->imp_unregistering); + } - DEBUG_REQ(D_INFO, req, "move req \"%s\" -> \"%s\"", - ptlrpc_rqphase2str(req), ptlrpc_phase2str(new_phase)); + DEBUG_REQ(D_INFO, req, "move req \"%s\" -> \"%s\"", + ptlrpc_rqphase2str(req), ptlrpc_phase2str(new_phase)); - req->rq_phase = new_phase; + req->rq_phase = new_phase; } /** - * Returns true if request \a req got early reply and hard deadline is not met + * Returns true if request \a req got early reply and hard deadline is not met */ static inline int ptlrpc_client_early(struct ptlrpc_request *req) @@ -3006,7 +3371,8 @@ ptlrpc_client_recv_or_unlink(struct ptlrpc_request *req) spin_unlock(&req->rq_lock); return 1; } - rc = req->rq_receiving_reply || req->rq_must_unlink; + rc = !req->rq_req_unlinked || !req->rq_reply_unlinked || + req->rq_receiving_reply; spin_unlock(&req->rq_lock); return rc; } @@ -3014,25 +3380,25 @@ ptlrpc_client_recv_or_unlink(struct ptlrpc_request *req) static inline void ptlrpc_client_wake_req(struct ptlrpc_request *req) { - if (req->rq_set == NULL) - cfs_waitq_signal(&req->rq_reply_waitq); - else - cfs_waitq_signal(&req->rq_set->set_waitq); + if (req->rq_set == NULL) + wake_up(&req->rq_reply_waitq); + else + wake_up(&req->rq_set->set_waitq); } static inline void ptlrpc_rs_addref(struct ptlrpc_reply_state *rs) { - LASSERT(cfs_atomic_read(&rs->rs_refcount) > 0); - cfs_atomic_inc(&rs->rs_refcount); + LASSERT(atomic_read(&rs->rs_refcount) > 0); + atomic_inc(&rs->rs_refcount); } static inline void ptlrpc_rs_decref(struct ptlrpc_reply_state *rs) { - LASSERT(cfs_atomic_read(&rs->rs_refcount) > 0); - if (cfs_atomic_dec_and_test(&rs->rs_refcount)) - lustre_free_reply_state(rs); + LASSERT(atomic_read(&rs->rs_refcount) > 0); + if (atomic_dec_and_test(&rs->rs_refcount)) + lustre_free_reply_state(rs); } /* Should only be called once per req */ @@ -3129,7 +3495,6 @@ int server_disconnect_export(struct obd_export *exp); * Pinger API (client side only) * @{ */ -extern int suppress_pings; enum timeout_event { TIMEOUT_GRANT = 1 }; @@ -3138,22 +3503,18 @@ typedef int (*timeout_cb_t)(struct timeout_item *, void *); int ptlrpc_pinger_add_import(struct obd_import *imp); int ptlrpc_pinger_del_import(struct obd_import *imp); int ptlrpc_add_timeout_client(int time, enum timeout_event event, - timeout_cb_t cb, void *data, - cfs_list_t *obd_list); -int ptlrpc_del_timeout_client(cfs_list_t *obd_list, + timeout_cb_t cb, void *data, + struct list_head *obd_list); +int ptlrpc_del_timeout_client(struct list_head *obd_list, enum timeout_event event); struct ptlrpc_request * ptlrpc_prep_ping(struct obd_import *imp); int ptlrpc_obd_ping(struct obd_device *obd); -cfs_time_t ptlrpc_suspend_wakeup_time(void); -#ifdef __KERNEL__ void ping_evictor_start(void); void ping_evictor_stop(void); -#else -#define ping_evictor_start() do {} while (0) -#define ping_evictor_stop() do {} while (0) -#endif -int ptlrpc_check_and_wait_suspend(struct ptlrpc_request *req); +void ptlrpc_pinger_ir_up(void); +void ptlrpc_pinger_ir_down(void); /** @} */ +int ptlrpc_pinger_suppress_pings(void); /* ptlrpc daemon bind policy */ typedef enum { @@ -3202,7 +3563,7 @@ void ptlrpcd_decref(void); * @{ */ const char* ll_opcode2str(__u32 opcode); -#ifdef LPROCFS +#ifdef CONFIG_PROC_FS void ptlrpc_lprocfs_register_obd(struct obd_device *obd); void ptlrpc_lprocfs_unregister_obd(struct obd_device *obd); void ptlrpc_lprocfs_brw(struct ptlrpc_request *req, int bytes); @@ -3220,7 +3581,6 @@ int llog_origin_handle_prev_block(struct ptlrpc_request *req); int llog_origin_handle_next_block(struct ptlrpc_request *req); int llog_origin_handle_read_header(struct ptlrpc_request *req); int llog_origin_handle_close(struct ptlrpc_request *req); -int llog_origin_handle_cancel(struct ptlrpc_request *req); /* ptlrpc/llog_client.c */ extern struct llog_operations llog_client_ops;