X-Git-Url: https://git.whamcloud.com/?p=fs%2Flustre-release.git;a=blobdiff_plain;f=lustre%2Finclude%2Flustre%2Flustre_idl.h;h=02c19d84f2c9589e9c6ee2b7bdbe9e6df47602df;hp=2a8786f998a62755a29a595b78002bb3e88cf6d1;hb=23628cd3fedf3f62b265245c4a9853da0330eff0;hpb=022b102258cd85314f4fa1fb8322638cc79f4634 diff --git a/lustre/include/lustre/lustre_idl.h b/lustre/include/lustre/lustre_idl.h index 2a8786f..02c19d8 100644 --- a/lustre/include/lustre/lustre_idl.h +++ b/lustre/include/lustre/lustre_idl.h @@ -1,10 +1,46 @@ /* -*- mode: c; c-basic-offset: 8; indent-tabs-mode: nil; -*- * vim:expandtab:shiftwidth=8:tabstop=8: * - * This file is part of Lustre, http://www.lustre.org + * GPL HEADER START + * + * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 only, + * as published by the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, but + * WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License version 2 for more details (a copy is included + * in the LICENSE file that accompanied this code). + * + * You should have received a copy of the GNU General Public License + * version 2 along with this program; If not, see + * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf + * + * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara, + * CA 95054 USA or visit www.sun.com if you need additional information or + * have any questions. + * + * GPL HEADER END + */ +/* + * Copyright 2008 Sun Microsystems, Inc. All rights reserved + * Use is subject to license terms. + */ +/* + * This file is part of Lustre, http://www.lustre.org/ + * Lustre is a trademark of Sun Microsystems, Inc. + * + * lustre/include/lustre/lustre_idl.h * * Lustre wire protocol definitions. + */ + +/** \defgroup lustreidl lustreidl * + * Lustre wire protocol definitions. * * We assume all nodes are either little-endian or big-endian, and we * always send messages in the sender's native format. The receiver @@ -47,25 +83,18 @@ * For variable length types, a second 'lustre_swab_v_xxxtypexxx()' routine * may be defined that swabs just the variable part, after the caller has * verified that the message buffer is large enough. + * + * @{ */ #ifndef _LUSTRE_IDL_H_ #define _LUSTRE_IDL_H_ -#include - -#if defined(__linux__) -#include -#elif defined(__APPLE__) -#include -#elif defined(__WINNT__) -#include -#else -#error Unsupported operating system. -#endif +#include /* Defn's shared with user-space. */ #include +#include /* * GENERAL STUFF @@ -193,55 +222,112 @@ static inline int range_is_exhausted(struct lu_range *r) (range)->lr_start, \ (range)->lr_end -struct lu_fid { - __u64 f_seq; /* holds fid sequence. Lustre should support 2 ^ 64 - * objects, thus even if one sequence has one object we - * reach this value. */ - __u32 f_oid; /* fid number within its sequence. */ - __u32 f_ver; /* holds fid version. */ -}; +/** \defgroup lu_fid lu_fid + * @{ */ -/* +/** + * File identifier. + * + * Fid is a cluster-wide unique identifier of a file or an object + * (stripe). Fids are never reused. Fids are transmitted across network (in + * the sender byte-ordering), and stored on disk in a packed form (struct + * lu_fid_pack) in a big-endian order. + */ +struct lu_fid { + /** + * fid sequence. Sequence is a unit of migration: all files (objects) + * with fids from a given sequence are stored on the same + * server. + * + * Lustre should support 2 ^ 64 objects, thus even if one + * sequence has one object we will never reach this value. + */ + __u64 f_seq; + /** fid number within sequence. */ + __u32 f_oid; + /** + * fid version, used to distinguish different versions (in the sense + * of snapshots, etc.) of the same file system object. Not currently + * used. + */ + __u32 f_ver; +}; + +/** * fid constants */ enum { - LUSTRE_ROOT_FID_SEQ = 1ULL, /* XXX: should go into mkfs. */ - /* initial fid id value */ LUSTRE_FID_INIT_OID = 1UL }; -/* get object sequence */ +/** returns fid object sequence */ static inline __u64 fid_seq(const struct lu_fid *fid) { return fid->f_seq; } -/* get object id */ +/** returns fid object id */ static inline __u32 fid_oid(const struct lu_fid *fid) { return fid->f_oid; } -/* get object version */ +/** returns fid object version */ static inline __u32 fid_ver(const struct lu_fid *fid) { return fid->f_ver; } -static inline int fid_seq_is_sane(__u64 seq) -{ - return seq != 0; -} - static inline void fid_zero(struct lu_fid *fid) { memset(fid, 0, sizeof(*fid)); } +/* Normal FID sequence starts from this value, i.e. 1<<33 */ +#define FID_SEQ_START 0x200000000ULL + +/* IDIF sequence starts from this value, i.e. 1<<32 */ +#define IDIF_SEQ_START 0x100000000ULL + +/** + * Check if a fid is igif or not. + * \param fid the fid to be tested. + * \return true if the fid is a igif; otherwise false. + */ static inline int fid_is_igif(const struct lu_fid *fid) { - return fid_seq(fid) == LUSTRE_ROOT_FID_SEQ; + return fid_seq(fid) > 0 && fid_seq(fid) < IDIF_SEQ_START; +} + +/** + * Check if a fid is idif or not. + * \param fid the fid to be tested. + * \return true if the fid is a idif; otherwise false. + */ +static inline int fid_is_idif(const struct lu_fid *fid) +{ + return fid_seq(fid) >= IDIF_SEQ_START && fid_seq(fid) < FID_SEQ_START; +} + +/** + * Get inode number from a igif. + * \param fid a igif to get inode number from. + * \return inode number for the igif. + */ +static inline ino_t lu_igif_ino(const struct lu_fid *fid) +{ + return fid_seq(fid); +} + +/** + * Get inode generation from a igif. + * \param fid a igif to get inode generation from. + * \return inode generation for the igif. + */ +static inline __u32 lu_igif_gen(const struct lu_fid *fid) +{ + return fid_oid(fid); } #define DFID "[0x%16.16"LPF64"x/0x%8.8x:0x%8.8x]" @@ -299,8 +385,7 @@ static inline void fid_be_to_cpu(struct lu_fid *dst, const struct lu_fid *src) LASSERTF(fid_is_igif(dst) || fid_ver(dst) == 0, DFID"\n", PFID(dst)); } -#ifdef __KERNEL__ -/* +/** * Storage representation for fids. * * Variable size, first byte contains the length of the whole record. @@ -315,14 +400,11 @@ void fid_pack(struct lu_fid_pack *pack, const struct lu_fid *fid, struct lu_fid *befider); int fid_unpack(const struct lu_fid_pack *pack, struct lu_fid *fid); -/* __KERNEL__ */ -#endif - static inline int fid_is_sane(const struct lu_fid *fid) { return fid != NULL && - ((fid_seq_is_sane(fid_seq(fid)) && fid_oid(fid) != 0 + ((fid_seq(fid) >= FID_SEQ_START && fid_oid(fid) != 0 && fid_ver(fid) == 0) || fid_is_igif(fid)); } @@ -346,21 +428,43 @@ static inline int lu_fid_eq(const struct lu_fid *f0, return memcmp(f0, f1, sizeof *f0) == 0; } -/* +#define __diff_normalize(val0, val1) \ +({ \ + typeof(val0) __val0 = (val0); \ + typeof(val1) __val1 = (val1); \ + \ + (__val0 == __val1 ? 0 : __val0 > __val1 ? +1 : -1); \ +}) + +static inline int lu_fid_cmp(const struct lu_fid *f0, + const struct lu_fid *f1) +{ + return + __diff_normalize(fid_seq(f0), fid_seq(f1)) ?: + __diff_normalize(fid_oid(f0), fid_oid(f1)) ?: + __diff_normalize(fid_ver(f0), fid_ver(f1)); +} + +/** @} lu_fid */ + +/** \defgroup lu_dir lu_dir + * @{ */ +/** * Layout of readdir pages, as transmitted on wire. */ struct lu_dirent { struct lu_fid lde_fid; - __u32 lde_hash; + __u64 lde_hash; __u16 lde_reclen; __u16 lde_namelen; + __u32 lde_pad0; char lde_name[0]; }; struct lu_dirpage { - __u32 ldp_hash_start; - __u32 ldp_hash_end; - __u16 ldp_flags; + __u64 ldp_hash_start; + __u64 ldp_hash_end; + __u32 ldp_flags; __u32 ldp_pad0; struct lu_dirent ldp_entries[0]; }; @@ -371,7 +475,7 @@ enum lu_dirpage_flags { static inline struct lu_dirent *lu_dirent_start(struct lu_dirpage *dp) { - if (le16_to_cpu(dp->ldp_flags) & LDF_EMPTY) + if (le32_to_cpu(dp->ldp_flags) & LDF_EMPTY) return NULL; else return dp->ldp_entries; @@ -393,12 +497,14 @@ static inline int lu_dirent_size(struct lu_dirent *ent) { if (le16_to_cpu(ent->lde_reclen) == 0) { return (sizeof(*ent) + - le16_to_cpu(ent->lde_namelen) + 3) & ~3; + le16_to_cpu(ent->lde_namelen) + 7) & ~7; } return le16_to_cpu(ent->lde_reclen); } -#define DIR_END_OFF 0xfffffffeUL +#define DIR_END_OFF 0xfffffffffffffffeULL + +/** @} lu_dir */ struct lustre_handle { __u64 cookie; @@ -422,23 +528,8 @@ static inline void lustre_handle_copy(struct lustre_handle *tgt, tgt->cookie = src->cookie; } -/* we depend on this structure to be 8-byte aligned */ -/* this type is only endian-adjusted in lustre_unpack_msg() */ -struct lustre_msg_v1 { - struct lustre_handle lm_handle; - __u32 lm_magic; - __u32 lm_type; - __u32 lm_version; - __u32 lm_opc; - __u64 lm_last_xid; - __u64 lm_last_committed; - __u64 lm_transno; - __u32 lm_status; - __u32 lm_flags; - __u32 lm_conn_cnt; - __u32 lm_bufcount; - __u32 lm_buflens[0]; -}; +/* flags for lm_flags */ +#define MSGHDR_AT_SUPPORT 0x1 #define lustre_msg lustre_msg_v2 /* we depend on this structure to be 8-byte aligned */ @@ -448,14 +539,15 @@ struct lustre_msg_v2 { __u32 lm_secflvr; __u32 lm_magic; __u32 lm_repsize; - __u32 lm_timeout; - __u32 lm_padding_1; + __u32 lm_cksum; + __u32 lm_flags; __u32 lm_padding_2; __u32 lm_padding_3; __u32 lm_buflens[0]; }; /* without gss, ptlrpc_body is put at the first buffer. */ +#define PTLRPC_NUM_VERSIONS 4 struct ptlrpc_body { struct lustre_handle pb_handle; __u32 pb_type; @@ -469,10 +561,14 @@ struct ptlrpc_body { __u32 pb_flags; __u32 pb_op_flags; __u32 pb_conn_cnt; - __u32 pb_padding_1; - __u32 pb_padding_2; + __u32 pb_timeout; /* for req, the deadline, for rep, the service est */ + __u32 pb_service_time; /* for rep, actual service time */ __u32 pb_limit; __u64 pb_slv; + /* VBR: pre-versions */ + __u64 pb_pre_versions[PTLRPC_NUM_VERSIONS]; + /* padding for future needs */ + __u64 pb_padding[4]; }; extern void lustre_swab_ptlrpc_body(struct ptlrpc_body *pb); @@ -497,17 +593,26 @@ extern void lustre_swab_ptlrpc_body(struct ptlrpc_body *pb); #define DLM_LOCKREPLY_OFF 1 /* lockrep offset */ #define DLM_REPLY_REC_OFF 2 /* reply record offset */ +/* only use in req->rq_{req,rep}_swab_mask */ +#define MSG_PTLRPC_HEADER_OFF 31 + /* Flags that are operation-specific go in the top 16 bits. */ #define MSG_OP_FLAG_MASK 0xffff0000 #define MSG_OP_FLAG_SHIFT 16 /* Flags that apply to all requests are in the bottom 16 bits */ -#define MSG_GEN_FLAG_MASK 0x0000ffff -#define MSG_LAST_REPLAY 1 -#define MSG_RESENT 2 -#define MSG_REPLAY 4 -#define MSG_REQ_REPLAY_DONE 8 -#define MSG_LOCK_REPLAY_DONE 16 +#define MSG_GEN_FLAG_MASK 0x0000ffff +#define MSG_LAST_REPLAY 0x0001 +#define MSG_RESENT 0x0002 +#define MSG_REPLAY 0x0004 +/* #define MSG_AT_SUPPORT 0x0008 + * This was used in early prototypes of adaptive timeouts, and while there + * shouldn't be any users of that code there also isn't a need for using this + * bits. Defer usage until at least 1.10 to avoid potential conflict. */ +#define MSG_DELAY_REPLAY 0x0010 +#define MSG_VERSION_REPLAY 0x0020 +#define MSG_REQ_REPLAY_DONE 0x0040 +#define MSG_LOCK_REPLAY_DONE 0x0080 /* * Flags for all connect opcodes (MDS_CONNECT, OST_CONNECT) @@ -549,10 +654,10 @@ extern void lustre_swab_ptlrpc_body(struct ptlrpc_body *pb); #define OBD_CONNECT_LRU_RESIZE 0x02000000ULL /* Lru resize feature. */ #define OBD_CONNECT_MDS_MDS 0x04000000ULL /* MDS-MDS connection*/ #define OBD_CONNECT_REAL 0x08000000ULL /* real connection */ -#define OBD_CONNECT_CHANGE_QS 0x10000000ULL /*shrink/enlarge qunit size - *b=10600 */ +#define OBD_CONNECT_CHANGE_QS 0x10000000ULL /* shrink/enlarge qunit b=10600 */ #define OBD_CONNECT_CKSUM 0x20000000ULL /* support several cksum algos */ #define OBD_CONNECT_FID 0x40000000ULL /* FID is supported by server */ +#define OBD_CONNECT_LOV_V3 0x100000000ULL /* client supports lov v3 ea */ /* also update obd_connect_names[] for lprocfs_rd_connect_flags() * and lustre/utils/wirecheck.c */ @@ -572,16 +677,17 @@ extern void lustre_swab_ptlrpc_body(struct ptlrpc_body *pb); OBD_CONNECT_MDS_CAPA | OBD_CONNECT_OSS_CAPA | \ OBD_CONNECT_MDS_MDS | OBD_CONNECT_CANCELSET | \ OBD_CONNECT_FID | \ - LRU_RESIZE_CONNECT_FLAG) + LRU_RESIZE_CONNECT_FLAG | OBD_CONNECT_AT | \ + OBD_CONNECT_LOV_V3) #define OST_CONNECT_SUPPORTED (OBD_CONNECT_SRVLOCK | OBD_CONNECT_GRANT | \ OBD_CONNECT_REQPORTAL | OBD_CONNECT_VERSION | \ OBD_CONNECT_TRUNCLOCK | OBD_CONNECT_INDEX | \ OBD_CONNECT_BRW_SIZE | OBD_CONNECT_QUOTA64 | \ OBD_CONNECT_OSS_CAPA | OBD_CONNECT_CANCELSET | \ - OBD_CONNECT_FID | \ - LRU_RESIZE_CONNECT_FLAG) + OBD_CONNECT_CKSUM | LRU_RESIZE_CONNECT_FLAG | \ + OBD_CONNECT_AT) #define ECHO_CONNECT_SUPPORTED (0) -#define MGS_CONNECT_SUPPORTED (OBD_CONNECT_VERSION | OBD_CONNECT_FID) +#define MGS_CONNECT_SUPPORTED (OBD_CONNECT_VERSION | OBD_CONNECT_AT) #define MAX_QUOTA_COUNT32 (0xffffffffULL) @@ -607,14 +713,25 @@ struct obd_connect_data { __u32 ocd_nllg; /* non-local-lustre-group */ __u64 ocd_transno; /* first transno from client to be replayed */ __u32 ocd_group; /* MDS group on OST */ - __u32 padding1; /* also fix lustre_swab_connect */ + __u32 ocd_cksum_types; /* supported checksum algorithms */ + __u64 padding1; /* also fix lustre_swab_connect */ __u64 padding2; /* also fix lustre_swab_connect */ - __u64 padding3; /* also fix lustre_swab_connect */ }; extern void lustre_swab_connect(struct obd_connect_data *ocd); /* + * Supported checksum algorithms. Up to 32 checksum types are supported. + * (32-bit mask stored in obd_connect_data::ocd_cksum_types) + * Please update DECLARE_CKSUM_NAME/OBD_CKSUM_ALL in obd.h when adding a new + * algorithm and also the OBD_FL_CKSUM* flags. + */ +typedef enum { + OBD_CKSUM_CRC32 = 0x00000001, + OBD_CKSUM_ADLER = 0x00000002, +} cksum_type_t; + +/* * OST requests: OBDO & OBD request records */ @@ -642,19 +759,19 @@ typedef enum { } ost_cmd_t; #define OST_FIRST_OPC OST_REPLY -typedef uint64_t obd_id; -typedef uint64_t obd_gr; -typedef uint64_t obd_time; -typedef uint64_t obd_size; -typedef uint64_t obd_off; -typedef uint64_t obd_blocks; -typedef uint32_t obd_blksize; -typedef uint32_t obd_mode; -typedef uint32_t obd_uid; -typedef uint32_t obd_gid; -typedef uint32_t obd_flag; -typedef uint64_t obd_valid; -typedef uint32_t obd_count; +typedef __u64 obd_id; +typedef __u64 obd_gr; +typedef __u64 obd_time; +typedef __u64 obd_size; +typedef __u64 obd_off; +typedef __u64 obd_blocks; +typedef __u32 obd_blksize; +typedef __u32 obd_mode; +typedef __u32 obd_uid; +typedef __u32 obd_gid; +typedef __u32 obd_flag; +typedef __u64 obd_valid; +typedef __u32 obd_count; #define OBD_FL_INLINEDATA (0x00000001) #define OBD_FL_OBDMDEXISTS (0x00000002) @@ -666,7 +783,7 @@ typedef uint32_t obd_count; #define OBD_FL_NO_USRQUOTA (0x00000100) /* the object's owner is over quota */ #define OBD_FL_NO_GRPQUOTA (0x00000200) /* the object's group is over quota */ -/* +/** * Set this to delegate DLM locking during obd_punch() to the OSTs. Only OSTs * that declared OBD_CONNECT_TRUNCLOCK in their connect flags support this * functionality. @@ -674,53 +791,16 @@ typedef uint32_t obd_count; #define OBD_FL_TRUNCLOCK (0x00000800) /* - * This should not be smaller than sizeof(struct lustre_handle) + sizeof(struct - * llog_cookie) + sizeof(struct ll_fid). Nevertheless struct ll_fid is not - * longer stored in o_inline, we keep this just for case. + * Checksum types */ -#define OBD_INLINESZ 80 - -/* Note: 64-bit types are 64-bit aligned in structure */ -struct obdo { - obd_valid o_valid; /* hot fields in this obdo */ - obd_id o_id; - obd_gr o_gr; - obd_id o_fid; - obd_size o_size; /* o_size-o_blocks == ost_lvb */ - obd_time o_mtime; - obd_time o_atime; - obd_time o_ctime; - obd_blocks o_blocks; /* brw: cli sent cached bytes */ - obd_size o_grant; - - /* 32-bit fields start here: keep an even number of them via padding */ - obd_blksize o_blksize; /* optimal IO blocksize */ - obd_mode o_mode; /* brw: cli sent cache remain */ - obd_uid o_uid; - obd_gid o_gid; - obd_flag o_flags; - obd_count o_nlink; /* brw: checksum */ - obd_count o_generation; - obd_count o_misc; /* brw: o_dropped */ - __u32 o_easize; /* epoch in ost writes */ - __u32 o_mds; - __u32 o_stripe_idx; /* holds stripe idx */ - __u32 o_padding_1; - char o_inline[OBD_INLINESZ]; - /* lustre_handle + llog_cookie */ -}; - -#define o_dirty o_blocks -#define o_undirty o_mode -#define o_dropped o_misc -#define o_cksum o_nlink - -extern void lustre_swab_obdo (struct obdo *o); - +#define OBD_FL_CKSUM_CRC32 (0x00001000) +#define OBD_FL_CKSUM_ADLER (0x00002000) +#define OBD_FL_CKSUM_ALL (OBD_FL_CKSUM_CRC32 | OBD_FL_CKSUM_ADLER) #define LOV_MAGIC_V1 0x0BD10BD0 #define LOV_MAGIC LOV_MAGIC_V1 #define LOV_MAGIC_JOIN 0x0BD20BD0 +#define LOV_MAGIC_V3 0x0BD30BD0 #define LOV_PATTERN_RAID0 0x001 /* stripes are used round-robin */ #define LOV_PATTERN_RAID1 0x002 /* stripes are mirrors of each other */ @@ -749,7 +829,7 @@ struct lov_mds_md_v1 { /* LOV EA mds/wire data (little-endian) */ struct lov_ost_data_v1 lmm_objects[0]; /* per-stripe data */ }; -extern void lustre_swab_lov_mds_md(struct lov_mds_md *llm); +/* extern void lustre_swab_lov_mds_md(struct lov_mds_md *llm); */ #define MAX_MD_SIZE (sizeof(struct lov_mds_md) + 4 * sizeof(struct lov_ost_data)) #define MIN_MD_SIZE (sizeof(struct lov_mds_md) + 1 * sizeof(struct lov_ost_data)) @@ -758,6 +838,18 @@ extern void lustre_swab_lov_mds_md(struct lov_mds_md *llm); #define XATTR_NAME_ACL_DEFAULT "system.posix_acl_default" #define XATTR_NAME_LOV "trusted.lov" +struct lov_mds_md_v3 { /* LOV EA mds/wire data (little-endian) */ + __u32 lmm_magic; /* magic number = LOV_MAGIC_V3 */ + __u32 lmm_pattern; /* LOV_PATTERN_RAID0, LOV_PATTERN_RAID1 */ + __u64 lmm_object_id; /* LOV object ID */ + __u64 lmm_object_gr; /* LOV object group */ + __u32 lmm_stripe_size; /* size of stripe in bytes */ + __u32 lmm_stripe_count; /* num stripes in use for this object */ + char lmm_pool_name[LOV_MAXPOOLNAME]; /* must be 32bit aligned */ + struct lov_ost_data_v1 lmm_objects[0]; /* per-stripe data */ +}; + + #define OBD_MD_FLID (0x00000001ULL) /* object ID */ #define OBD_MD_FLATIME (0x00000002ULL) /* access time */ #define OBD_MD_FLMTIME (0x00000004ULL) /* data modification time */ @@ -772,11 +864,11 @@ extern void lustre_swab_lov_mds_md(struct lov_mds_md *llm); #define OBD_MD_FLFLAGS (0x00000800ULL) /* flags word */ #define OBD_MD_FLNLINK (0x00002000ULL) /* link count */ #define OBD_MD_FLGENER (0x00004000ULL) /* generation number */ -#define OBD_MD_FLINLINE (0x00008000ULL) /* inline data */ +/*#define OBD_MD_FLINLINE (0x00008000ULL) inline data. used until 1.6.5 */ #define OBD_MD_FLRDEV (0x00010000ULL) /* device number */ #define OBD_MD_FLEASIZE (0x00020000ULL) /* extended attribute data */ #define OBD_MD_LINKNAME (0x00040000ULL) /* symbolic link target */ -#define OBD_MD_FLHANDLE (0x00080000ULL) /* file handle */ +#define OBD_MD_FLHANDLE (0x00080000ULL) /* file/lock handle */ #define OBD_MD_FLCKSUM (0x00100000ULL) /* bulk data checksum */ #define OBD_MD_FLQOS (0x00200000ULL) /* quality of service stats */ #define OBD_MD_FLOSCOPQ (0x00400000ULL) /* osc opaque data */ @@ -816,16 +908,6 @@ extern void lustre_swab_lov_mds_md(struct lov_mds_md *llm); OBD_MD_FLGID | OBD_MD_FLFLAGS | OBD_MD_FLNLINK | \ OBD_MD_FLGENER | OBD_MD_FLRDEV | OBD_MD_FLGROUP) -static inline struct lustre_handle *obdo_handle(struct obdo *oa) -{ - return (struct lustre_handle *)oa->o_inline; -} - -static inline struct llog_cookie *obdo_logcookie(struct obdo *oa) -{ - return (struct llog_cookie *)(oa->o_inline + - sizeof(struct lustre_handle)); -} /* don't forget obdo_fid which is way down at the bottom so it can * come after the definition of llog_cookie */ @@ -853,6 +935,11 @@ struct obd_statfs { }; extern void lustre_swab_obd_statfs (struct obd_statfs *os); +#define OBD_STATFS_NODELAY 0x0001 /* requests should be send without delay + * and resends for avoid deadlocks */ + +#define OBD_STATFS_FROM_CACHE 0x0002 /* the statfs callback should not update + * obd_osfs_age */ /* ost_body.data values for OST_BRW */ @@ -890,15 +977,6 @@ struct niobuf_remote { extern void lustre_swab_niobuf_remote (struct niobuf_remote *nbr); -/* request structure for OST's */ - -struct ost_body { - struct obdo oa; -}; - -extern void lustre_swab_ost_body (struct ost_body *b); -extern void lustre_swab_ost_last_id(obd_id *id); - /* lock value block communicated between the filter and llite */ /* OST_LVB_ERR_INIT is needed because the return code in rc is @@ -1129,7 +1207,7 @@ struct mds_body { __u32 eadatasize; __u32 aclsize; __u32 max_mdsize; - __u32 max_cookiesize; /* also fix lustre_swab_mds_body */ + __u32 max_cookiesize; __u32 padding_4; /* also fix lustre_swab_mds_body */ }; @@ -1321,7 +1399,8 @@ enum { MDS_CHECK_SPLIT = 1 << 0, MDS_CROSS_REF = 1 << 1, MDS_VTX_BYPASS = 1 << 2, - MDS_PERM_BYPASS = 1 << 3 + MDS_PERM_BYPASS = 1 << 3, + MDS_SOM = 1 << 4 }; struct mds_rec_join { @@ -1515,7 +1594,7 @@ struct mdt_rec_setxattr { __u32 sx_padding_2; __u32 sx_padding_3; __u64 sx_valid; - __u64 sx_padding_4; + __u64 sx_time; __u64 sx_padding_5; __u64 sx_padding_6; __u64 sx_padding_7; @@ -1813,12 +1892,15 @@ struct cfg_marker { __u32 cm_flags; __u32 cm_vers; /* lustre release version number */ __u32 padding; /* 64 bit align */ - time_t cm_createtime; /*when this record was first created */ - time_t cm_canceltime; /*when this record is no longer valid*/ + __u64 cm_createtime; /*when this record was first created */ + __u64 cm_canceltime; /*when this record is no longer valid*/ char cm_tgtname[MTI_NAME_MAXLEN]; char cm_comment[MTI_NAME_MAXLEN]; }; +extern void lustre_swab_cfg_marker(struct cfg_marker *marker, + int swab, int size); + /* * Opcodes for multiple servers. */ @@ -1833,14 +1915,14 @@ typedef enum { /* catalog of log objects */ -/* Identifier for a single log object */ +/** Identifier for a single log object */ struct llog_logid { __u64 lgl_oid; __u64 lgl_ogr; __u32 lgl_ogen; } __attribute__((packed)); -/* Records written to the CATALOGS list */ +/** Records written to the CATALOGS list */ #define CATLIST "CATALOGS" struct llog_catid { struct llog_logid lci_logid; @@ -1849,7 +1931,7 @@ struct llog_catid { __u32 lci_padding3; } __attribute__((packed)); -/*join file lov mds md*/ +/** join file lov mds md*/ struct lov_mds_md_join { struct lov_mds_md lmmj_md; /*join private info*/ @@ -1886,7 +1968,7 @@ typedef enum { __swab32(LLOG_OP_MAGIC) || \ (((r)->lrh_type == 0) && ((r)->lrh_len > LLOG_CHUNK_SIZE))) -/* Log record header - stored in little endian order. +/** Log record header - stored in little endian order. * Each record must start with this struct, end with a llog_rec_tail, * and be a multiple of 256 bits in size. */ @@ -1913,7 +1995,7 @@ struct llog_logid_rec { struct llog_rec_tail lid_tail; } __attribute__((packed)); -/* MDS extent description +/** MDS extent description * It is for joined file extent info, each extent info for joined file * just like (start, end, lmm). */ @@ -1922,7 +2004,8 @@ struct mds_extent_desc { __u64 med_len; /* extent length */ struct lov_mds_md med_lmm; /* extent's lmm */ }; -/*Joined file array extent log record*/ + +/** Joined file array extent log record*/ struct llog_array_rec { struct llog_rec_hdr lmr_hdr; struct mds_extent_desc lmr_med; @@ -2013,7 +2096,7 @@ struct llog_log_hdr { llh->llh_bitmap_offset - \ sizeof(llh->llh_tail)) * 8) -/* log cookies are used to reference a specific log file and a record therein */ +/** log cookies are used to reference a specific log file and a record therein */ struct llog_cookie { struct llog_logid lgc_lgl; __u32 lgc_subsys; @@ -2021,7 +2104,7 @@ struct llog_cookie { __u32 lgc_padding; } __attribute__((packed)); -/* llog protocol */ +/** llog protocol */ enum llogd_rpc_ops { LLOG_ORIGIN_HANDLE_CREATE = 501, LLOG_ORIGIN_HANDLE_NEXT_BLOCK = 502, @@ -2075,8 +2158,70 @@ struct lov_user_md_join { /* LOV EA user data (host-endian) */ struct lov_user_ost_data_join lmm_objects[0]; /* per-stripe data */ } __attribute__((packed)); -extern void lustre_swab_lov_user_md(struct lov_user_md *lum); -extern void lustre_swab_lov_user_md_objects(struct lov_user_md *lum); +/* Note: 64-bit types are 64-bit aligned in structure */ +struct obdo { + obd_valid o_valid; /* hot fields in this obdo */ + obd_id o_id; + obd_gr o_gr; + obd_id o_fid; + obd_size o_size; /* o_size-o_blocks == ost_lvb */ + obd_time o_mtime; + obd_time o_atime; + obd_time o_ctime; + obd_blocks o_blocks; /* brw: cli sent cached bytes */ + obd_size o_grant; + + /* 32-bit fields start here: keep an even number of them via padding */ + obd_blksize o_blksize; /* optimal IO blocksize */ + obd_mode o_mode; /* brw: cli sent cache remain */ + obd_uid o_uid; + obd_gid o_gid; + obd_flag o_flags; + obd_count o_nlink; /* brw: checksum */ + obd_count o_generation; + obd_count o_misc; /* brw: o_dropped */ + __u32 o_easize; /* epoch in ost writes */ + __u32 o_mds; + __u32 o_stripe_idx; /* holds stripe idx */ + __u32 o_padding_1; + struct lustre_handle o_handle; /* brw: lock handle to prolong locks */ + struct llog_cookie o_lcookie; /* destroy: unlink cookie from MDS */ + + __u64 o_padding_2; + __u64 o_padding_3; + __u64 o_padding_4; + __u64 o_padding_5; + __u64 o_padding_6; +}; + +#define o_dirty o_blocks +#define o_undirty o_mode +#define o_dropped o_misc +#define o_cksum o_nlink + +extern void lustre_swab_obdo (struct obdo *o); + +/* request structure for OST's */ + +struct ost_body { + struct obdo oa; +}; + +/* Key for FIEMAP to be used in get_info calls */ +struct ll_fiemap_info_key { + char name[8]; + struct obdo oa; + struct ll_user_fiemap fiemap; +}; + +extern void lustre_swab_ost_body (struct ost_body *b); +extern void lustre_swab_ost_last_id(obd_id *id); +extern void lustre_swab_fiemap(struct ll_user_fiemap *fiemap); + +extern void lustre_swab_lov_user_md_v1(struct lov_user_md_v1 *lum); +extern void lustre_swab_lov_user_md_v3(struct lov_user_md_v3 *lum); +extern void lustre_swab_lov_user_md_objects(struct lov_user_ost_data *lod, + int stripe_count); extern void lustre_swab_lov_user_md_join(struct lov_user_md_join *lumj); /* llog_swab.c */ @@ -2119,7 +2264,7 @@ typedef enum { #define JOIN_FILE_ALIGN 4096 -/* security opcodes */ +/** security opcodes */ typedef enum { SEC_CTX_INIT = 801, SEC_CTX_INIT_CONT = 802, @@ -2150,18 +2295,18 @@ struct lustre_capa { extern void lustre_swab_lustre_capa(struct lustre_capa *c); -/* lustre_capa.lc_opc */ +/** lustre_capa::lc_opc */ enum { - CAPA_OPC_BODY_WRITE = 1<<0, /* write object data */ - CAPA_OPC_BODY_READ = 1<<1, /* read object data */ - CAPA_OPC_INDEX_LOOKUP = 1<<2, /* lookup object fid */ - CAPA_OPC_INDEX_INSERT = 1<<3, /* insert object fid */ - CAPA_OPC_INDEX_DELETE = 1<<4, /* delete object fid */ - CAPA_OPC_OSS_WRITE = 1<<5, /* write oss object data */ - CAPA_OPC_OSS_READ = 1<<6, /* read oss object data */ - CAPA_OPC_OSS_TRUNC = 1<<7, /* truncate oss object */ - CAPA_OPC_META_WRITE = 1<<8, /* write object meta data */ - CAPA_OPC_META_READ = 1<<9, /* read object meta data */ + CAPA_OPC_BODY_WRITE = 1<<0, /**< write object data */ + CAPA_OPC_BODY_READ = 1<<1, /**< read object data */ + CAPA_OPC_INDEX_LOOKUP = 1<<2, /**< lookup object fid */ + CAPA_OPC_INDEX_INSERT = 1<<3, /**< insert object fid */ + CAPA_OPC_INDEX_DELETE = 1<<4, /**< delete object fid */ + CAPA_OPC_OSS_WRITE = 1<<5, /**< write oss object data */ + CAPA_OPC_OSS_READ = 1<<6, /**< read oss object data */ + CAPA_OPC_OSS_TRUNC = 1<<7, /**< truncate oss object */ + CAPA_OPC_META_WRITE = 1<<8, /**< write object meta data */ + CAPA_OPC_META_READ = 1<<9, /**< read object meta data */ }; @@ -2189,9 +2334,9 @@ static inline int capa_for_oss(struct lustre_capa *c) return (c->lc_opc & CAPA_OPC_INDEX_LOOKUP) == 0; } -/* lustre_capa.lc_hmac_alg */ +/* lustre_capa::lc_hmac_alg */ enum { - CAPA_HMAC_ALG_SHA1 = 1, /* sha1 algorithm */ + CAPA_HMAC_ALG_SHA1 = 1, /**< sha1 algorithm */ CAPA_HMAC_ALG_MAX, }; @@ -2199,17 +2344,19 @@ enum { #define CAPA_HMAC_ALG_MASK 0xff000000 struct lustre_capa_key { - __u64 lk_mdsid; /* mds# */ - __u32 lk_keyid; /* key# */ + __u64 lk_mdsid; /**< mds# */ + __u32 lk_keyid; /**< key# */ __u32 lk_padding; - __u8 lk_key[CAPA_HMAC_KEY_MAX_LEN]; /* key */ + __u8 lk_key[CAPA_HMAC_KEY_MAX_LEN]; /**< key */ } __attribute__((packed)); extern void lustre_swab_lustre_capa_key(struct lustre_capa_key *k); /* quota check function */ -#define QUOTA_RET_OK 0 /* return successfully */ -#define QUOTA_RET_NOQUOTA 1 /* not support quota */ -#define QUOTA_RET_NOLIMIT 2 /* quota limit isn't set */ -#define QUOTA_RET_ACQUOTA 3 /* need to acquire extra quota */ +#define QUOTA_RET_OK 0 /**< return successfully */ +#define QUOTA_RET_NOQUOTA 1 /**< not support quota */ +#define QUOTA_RET_NOLIMIT 2 /**< quota limit isn't set */ +#define QUOTA_RET_ACQUOTA 3 /**< need to acquire extra quota */ #endif + +/** @} lustreidl */