X-Git-Url: https://git.whamcloud.com/?a=blobdiff_plain;f=lustre%2Finclude%2Flustre%2Flustre_idl.h;h=bae42d0dc06006eb3d865e9ef1ad9c2cbcfd81fa;hb=5b64ac7f7cf2767acb75b872eaffcf6d255d0501;hp=4c854cfe5fbbb2c0128676ee92dde13d3a622d4d;hpb=8a06a00535b929a8b5761764457be03a3308f6ec;p=fs%2Flustre-release.git diff --git a/lustre/include/lustre/lustre_idl.h b/lustre/include/lustre/lustre_idl.h index 4c854cf..bae42d0 100644 --- a/lustre/include/lustre/lustre_idl.h +++ b/lustre/include/lustre/lustre_idl.h @@ -418,9 +418,12 @@ enum fid_seq { FID_SEQ_IDIF_MAX = 0x1ffffffffULL, /* Normal FID sequence starts from this value, i.e. 1<<33 */ FID_SEQ_START = 0x200000000ULL, + /* sequence for local pre-defined FIDs listed in local_oid */ FID_SEQ_LOCAL_FILE = 0x200000001ULL, FID_SEQ_DOT_LUSTRE = 0x200000002ULL, - /* XXX 0x200000003ULL is reserved for FID_SEQ_LLOG_OBJ */ + /* sequence is used for local named objects FIDs generated + * by local_object_storage library */ + FID_SEQ_LOCAL_NAME = 0x200000003ULL, FID_SEQ_SPECIAL = 0x200000004ULL, FID_SEQ_QUOTA = 0x200000005ULL, FID_SEQ_QUOTA_GLB = 0x200000006ULL, @@ -530,6 +533,13 @@ static inline obd_id fid_idif_id(obd_seq seq, __u32 oid, __u32 ver) return ((__u64)ver << 48) | ((seq & 0xffff) << 32) | oid; } +/* extract ost index from IDIF FID */ +static inline __u32 fid_idif_ost_idx(const struct lu_fid *fid) +{ + LASSERT(fid_is_idif(fid)); + return (fid_seq(fid) >> 16) & 0xffff; +} + /* unpack an ostid (id/seq) from a wire/disk structure into an IDIF FID */ static inline void ostid_idif_unpack(struct ost_id *ostid, struct lu_fid *fid, __u32 ost_idx) @@ -948,7 +958,7 @@ static inline int lu_dirent_size(struct lu_dirent *ent) #define LU_PAGE_SIZE (1UL << LU_PAGE_SHIFT) #define LU_PAGE_MASK (~(LU_PAGE_SIZE - 1)) -#define LU_PAGE_COUNT 1 << (CFS_PAGE_SHIFT - LU_PAGE_SHIFT) +#define LU_PAGE_COUNT (1 << (CFS_PAGE_SHIFT - LU_PAGE_SHIFT)) /** @} lu_dir */ @@ -1152,9 +1162,12 @@ extern void lustre_swab_ptlrpc_body(struct ptlrpc_body *pb); #define OBD_CONNECT_JOBSTATS 0x20000000000ULL /* jobid in ptlrpc_body */ #define OBD_CONNECT_UMASK 0x40000000000ULL /* create uses client umask */ #define OBD_CONNECT_EINPROGRESS 0x80000000000ULL /* client handles -EINPROGRESS - * write RPC error properly */ + * RPC error properly */ #define OBD_CONNECT_GRANT_PARAM 0x100000000000ULL/* extra grant params used for * finer space reservation */ +#define OBD_CONNECT_NANOSEC_TIME 0x200000000000ULL /* nanosecond timestamps */ +#define OBD_CONNECT_LVB_TYPE 0x400000000000ULL /* variable type of LVB */ +#define OBD_CONNECT_LIGHTWEIGHT 0x1000000000000ULL/* lightweight connection */ /* XXX README XXX: * Please DO NOT add flag values here before first ensuring that this same * flag value is not in use on some other branch. Please clear any such @@ -1163,6 +1176,12 @@ extern void lustre_swab_ptlrpc_body(struct ptlrpc_body *pb); * and updates obd_connect_names[] for lprocfs_rd_connect_flags(), so it * can be approved and landed easily to reserve the flag for future use. */ +/* The MNE_SWAB flag is overloading the MDS_MDS bit only for the MGS + * connection. It is a temporary bug fix for Imperative Recovery interop + * between 2.2 and 2.3 x86/ppc nodes, and can be removed when interop for + * 2.2 clients/servers is no longer needed. LU-1252/LU-1644. */ +#define OBD_CONNECT_MNE_SWAB OBD_CONNECT_MDS_MDS + #define OCD_HAS_FLAG(ocd, flg) \ (!!((ocd)->ocd_connect_flags & OBD_CONNECT_##flg)) @@ -1185,8 +1204,9 @@ extern void lustre_swab_ptlrpc_body(struct ptlrpc_body *pb); OBD_CONNECT_FID | LRU_RESIZE_CONNECT_FLAG | \ OBD_CONNECT_VBR | OBD_CONNECT_LOV_V3 | \ OBD_CONNECT_SOM | OBD_CONNECT_FULL20 | \ - OBD_CONNECT_64BITHASH | \ - OBD_CONNECT_EINPROGRESS | OBD_CONNECT_JOBSTATS) + OBD_CONNECT_64BITHASH | OBD_CONNECT_JOBSTATS | \ + OBD_CONNECT_EINPROGRESS | \ + OBD_CONNECT_LIGHTWEIGHT) #define OST_CONNECT_SUPPORTED (OBD_CONNECT_SRVLOCK | OBD_CONNECT_GRANT | \ OBD_CONNECT_REQPORTAL | OBD_CONNECT_VERSION | \ OBD_CONNECT_TRUNCLOCK | OBD_CONNECT_INDEX | \ @@ -1201,10 +1221,12 @@ extern void lustre_swab_ptlrpc_body(struct ptlrpc_body *pb); OBD_CONNECT_GRANT_SHRINK | OBD_CONNECT_FULL20 | \ OBD_CONNECT_64BITHASH | OBD_CONNECT_MAXBYTES | \ OBD_CONNECT_MAX_EASIZE | \ - OBD_CONNECT_EINPROGRESS | OBD_CONNECT_JOBSTATS) + OBD_CONNECT_EINPROGRESS | \ + OBD_CONNECT_JOBSTATS | OBD_CONNECT_LIGHTWEIGHT) #define ECHO_CONNECT_SUPPORTED (0) #define MGS_CONNECT_SUPPORTED (OBD_CONNECT_VERSION | OBD_CONNECT_AT | \ - OBD_CONNECT_FULL20 | OBD_CONNECT_IMP_RECOV) + OBD_CONNECT_FULL20 | OBD_CONNECT_IMP_RECOV | \ + OBD_CONNECT_MNE_SWAB) /* Features required for this version of the client to work with server */ #define CLIENT_CONNECT_MDT_REQD (OBD_CONNECT_IBITS | OBD_CONNECT_FID | \ @@ -1367,6 +1389,21 @@ enum obdo_flags { #define LOV_MAGIC_JOIN_V1 0x0BD20BD0 #define LOV_MAGIC_V3 0x0BD30BD0 +/* + * magic for fully defined striping + * the idea is that we should have different magics for striping "hints" + * (struct lov_user_md_v[13]) and defined ready-to-use striping (struct + * lov_mds_md_v[13]). at the moment the magics are used in wire protocol, + * we can't just change it w/o long way preparation, but we still need a + * mechanism to allow LOD to differentiate hint versus ready striping. + * so, at the moment we do a trick: MDT knows what to expect from request + * depending on the case (replay uses ready striping, non-replay req uses + * hints), so MDT replaces magic with appropriate one and now LOD can + * easily understand what's inside -bzzz + */ +#define LOV_MAGIC_V1_DEF 0x0CD10BD0 +#define LOV_MAGIC_V3_DEF 0x0CD30BD0 + #define LOV_PATTERN_RAID0 0x001 /* stripes are used round-robin */ #define LOV_PATTERN_RAID1 0x002 /* stripes are mirrors of each other */ #define LOV_PATTERN_FIRST 0x100 /* first stripe is not in round-robin */ @@ -1513,6 +1550,8 @@ extern void lustre_swab_obd_statfs (struct obd_statfs *os); #define OBD_BRW_SRVLOCK 0x200 /* Client holds no lock over this page */ #define OBD_BRW_ASYNC 0x400 /* Server may delay commit to disk */ #define OBD_BRW_MEMALLOC 0x800 /* Client runs in the "kswapd" context */ +#define OBD_BRW_OVER_USRQUOTA 0x1000 /* Running out of user quota */ +#define OBD_BRW_OVER_GRPQUOTA 0x2000 /* Running out of group quota */ #define OBD_OBJECT_EOF 0xffffffffffffffffULL @@ -1642,6 +1681,8 @@ enum { LUSTRE_RES_ID_SEQ_OFF = 0, LUSTRE_RES_ID_VER_OID_OFF = 1, LUSTRE_RES_ID_WAS_VER_OFF = 2, /* see note above */ + LUSTRE_RES_ID_QUOTA_SEQ_OFF = 2, + LUSTRE_RES_ID_QUOTA_VER_OID_OFF = 3, LUSTRE_RES_ID_HSH_OFF = 3 }; @@ -1755,6 +1796,17 @@ struct mdt_ioepoch { extern void lustre_swab_mdt_ioepoch (struct mdt_ioepoch *b); +/* The lquota_id structure is an union of all the possible identifier types that + * can be used with quota, this includes: + * - 64-bit user ID + * - 64-bit group ID + * - a FID which can be used for per-directory quota in the future */ +union lquota_id { + struct lu_fid qid_fid; /* FID for per-directory quota */ + __u64 qid_uid; /* user identifier */ + __u64 qid_gid; /* group identifier */ +}; + #define Q_QUOTACHECK 0x800100 #define Q_INITQUOTA 0x800101 /* init slave limits */ #define Q_GETOINFO 0x800102 /* get obd quota info */ @@ -1790,6 +1842,35 @@ struct obd_quotactl { extern void lustre_swab_obd_quotactl(struct obd_quotactl *q); +#define QUOTA_DQACQ_FL_ACQ 0x1 /* acquire quota */ +#define QUOTA_DQACQ_FL_PREACQ 0x2 /* pre-acquire */ +#define QUOTA_DQACQ_FL_REL 0x4 /* release quota */ +#define QUOTA_DQACQ_FL_REPORT 0x8 /* report usage */ + +struct quota_body { + struct lu_fid qb_fid; /* FID of global index packing the pool ID + * and type (data or metadata) as well as + * the quota type (user or group). */ + union lquota_id qb_id; /* uid or gid or directory FID */ + __u32 qb_flags; /* see above */ + __u32 qb_padding; + __u64 qb_count; /* acquire/release count (kbytes/inodes) */ + __u64 qb_usage; /* current slave usage (kbytes/inodes) */ + __u64 qb_slv_ver; /* slave index file version */ + struct lustre_handle qb_lockh; /* per-ID lock handle */ + struct lustre_handle qb_glb_lockh; /* global lock handle */ + __u64 qb_padding1[4]; +}; + +/* When the quota_body is used in the reply of quota global intent + * lock (IT_QUOTA_CONN) reply, qb_fid contains slave index file FID. */ +#define qb_slv_fid qb_fid +/* qb_usage is the current qunit (in kbytes/inodes) when quota_body is used in + * quota reply */ +#define qb_qunit qb_usage + +extern void lustre_swab_quota_body(struct quota_body *b); + struct quota_adjust_qunit { __u32 qaq_flags; __u32 qaq_id; @@ -1799,6 +1880,57 @@ struct quota_adjust_qunit { }; extern void lustre_swab_quota_adjust_qunit(struct quota_adjust_qunit *q); +/* Quota types currently supported */ +enum { + LQUOTA_TYPE_USR = 0x00, /* maps to USRQUOTA */ + LQUOTA_TYPE_GRP = 0x01, /* maps to GRPQUOTA */ + LQUOTA_TYPE_MAX +}; + +/* There are 2 different resource types on which a quota limit can be enforced: + * - inodes on the MDTs + * - blocks on the OSTs */ +enum { + LQUOTA_RES_MD = 0x01, /* skip 0 to avoid null oid in FID */ + LQUOTA_RES_DT = 0x02, + LQUOTA_LAST_RES, + LQUOTA_FIRST_RES = LQUOTA_RES_MD +}; +#define LQUOTA_NR_RES (LQUOTA_LAST_RES - LQUOTA_FIRST_RES + 1) + +/* + * Space accounting support + * Format of an accounting record, providing disk usage information for a given + * user or group + */ +struct lquota_acct_rec { /* 16 bytes */ + __u64 bspace; /* current space in use */ + __u64 ispace; /* current # inodes in use */ +}; + +/* + * Global quota index support + * Format of a global record, providing global quota settings for a given quota + * identifier + */ +struct lquota_glb_rec { /* 32 bytes */ + __u64 qbr_hardlimit; /* quota hard limit, in #inodes or kbytes */ + __u64 qbr_softlimit; /* quota soft limit, in #inodes or kbytes */ + __u64 qbr_time; /* grace time, in seconds */ + __u64 qbr_granted; /* how much is granted to slaves, in #inodes or + * kbytes */ +}; + +/* + * Slave index support + * Format of a slave record, recording how much space is granted to a given + * slave + */ +struct lquota_slv_rec { /* 8 bytes */ + __u64 qsr_granted; /* space granted to the slave for the key=ID, + * in #inodes or kbytes */ +}; + /* flags is shared among quota structures */ #define LQUOTA_FLAGS_GRP 1UL /* 0 is user, 1 is group */ #define LQUOTA_FLAGS_BLK 2UL /* 0 is inode, 1 is block */ @@ -1926,6 +2058,7 @@ extern void lustre_swab_mdt_rec_setattr (struct mdt_rec_setattr *sa); #define MDS_OPEN_SYNC 00010000 #define MDS_OPEN_DIRECTORY 00200000 +#define MDS_OPEN_BY_FID 040000000 /* open_by_fid for known object */ #define MDS_OPEN_DELAY_CREATE 0100000000 /* delay initial object create */ #define MDS_OPEN_OWNEROVERRIDE 0200000000 /* NFSD rw-reopen ro file for owner */ #define MDS_OPEN_JOIN_FILE 0400000000 /* open for join file. @@ -1972,7 +2105,6 @@ enum { MDS_CLOSE_CLEANUP = 1 << 6, MDS_KEEP_ORPHAN = 1 << 7, MDS_RECOV_OPEN = 1 << 8, - MDS_UNLINK_DESTROY = 1 << 9, /* Destory ost object in mdd_unlink */ }; /* instance of mdt_reint_rec */ @@ -2346,15 +2478,52 @@ typedef union { extern void lustre_swab_ldlm_policy_data (ldlm_wire_policy_data_t *d); +/* Data structures associated with the quota locks */ + +/* Glimpse descriptor used for the index & per-ID quota locks */ +struct ldlm_gl_lquota_desc { + union lquota_id gl_id; /* quota ID subject to the glimpse */ + __u64 gl_flags; /* see LQUOTA_FL* below */ + __u64 gl_ver; /* new index version */ + __u64 gl_hardlimit; /* new hardlimit or qunit value */ + __u64 gl_softlimit; /* new softlimit */ + __u64 gl_pad1; + __u64 gl_pad2; +}; +#define gl_qunit gl_hardlimit /* current qunit value used when + * glimpsing per-ID quota locks */ + +/* quota glimpse flags */ +#define LQUOTA_FL_EDQUOT 0x1 /* user/group out of quota space on QMT */ + +/* LVB used with quota (global and per-ID) locks */ +struct lquota_lvb { + __u64 lvb_flags; /* see LQUOTA_FL* above */ + __u64 lvb_id_may_rel; /* space that might be released later */ + __u64 lvb_id_rel; /* space released by the slave for this ID */ + __u64 lvb_id_qunit; /* current qunit value */ + __u64 lvb_pad1; +}; + +/* LVB used with global quota lock */ +#define lvb_glb_ver lvb_id_may_rel /* current version of the global index */ + /* Similarly to ldlm_wire_policy_data_t, there is one common swabber for all * LVB types. As a result, any new LVB structure must match the fields of the * ost_lvb structure. */ union ldlm_wire_lvb { - struct ost_lvb l_ost; + struct ost_lvb l_ost; + struct lquota_lvb l_lquota; }; extern void lustre_swab_lvb(union ldlm_wire_lvb *); +union ldlm_gl_desc { + struct ldlm_gl_lquota_desc lquota_desc; +}; + +extern void lustre_swab_gl_desc(union ldlm_gl_desc *); + struct ldlm_intent { __u64 opc; }; @@ -2514,6 +2683,7 @@ typedef enum { OBD_PING = 400, OBD_LOG_CANCEL, OBD_QC_CALLBACK, + OBD_IDX_READ, OBD_LAST_OPC } obd_cmd_t; #define OBD_FIRST_OPC OBD_PING @@ -2543,111 +2713,92 @@ struct llog_catid { #define LLOG_OP_MASK 0xfff00000 typedef enum { - LLOG_PAD_MAGIC = LLOG_OP_MAGIC | 0x00000, - OST_SZ_REC = LLOG_OP_MAGIC | 0x00f00, - OST_RAID1_REC = LLOG_OP_MAGIC | 0x01000, - MDS_UNLINK_REC = LLOG_OP_MAGIC | 0x10000 | (MDS_REINT << 8) | REINT_UNLINK, - MDS_SETATTR_REC = LLOG_OP_MAGIC | 0x10000 | (MDS_REINT << 8) | REINT_SETATTR, - MDS_SETATTR64_REC = LLOG_OP_MAGIC | 0x90000 | (MDS_REINT << 8) | REINT_SETATTR, - OBD_CFG_REC = LLOG_OP_MAGIC | 0x20000, - PTL_CFG_REC = LLOG_OP_MAGIC | 0x30000, /* obsolete */ - LLOG_GEN_REC = LLOG_OP_MAGIC | 0x40000, - LLOG_JOIN_REC = LLOG_OP_MAGIC | 0x50000, /* obsolete */ - CHANGELOG_REC = LLOG_OP_MAGIC | 0x60000, - CHANGELOG_USER_REC = LLOG_OP_MAGIC | 0x70000, - LLOG_HDR_MAGIC = LLOG_OP_MAGIC | 0x45539, - LLOG_LOGID_MAGIC = LLOG_OP_MAGIC | 0x4553b, + LLOG_PAD_MAGIC = LLOG_OP_MAGIC | 0x00000, + OST_SZ_REC = LLOG_OP_MAGIC | 0x00f00, + /* OST_RAID1_REC = LLOG_OP_MAGIC | 0x01000, never used */ + MDS_UNLINK_REC = LLOG_OP_MAGIC | 0x10000 | (MDS_REINT << 8) | + REINT_UNLINK, /* obsolete after 2.5.0 */ + MDS_UNLINK64_REC = LLOG_OP_MAGIC | 0x90000 | (MDS_REINT << 8) | + REINT_UNLINK, + /* MDS_SETATTR_REC = LLOG_OP_MAGIC | 0x12401, obsolete 1.8.0 */ + MDS_SETATTR64_REC = LLOG_OP_MAGIC | 0x90000 | (MDS_REINT << 8) | + REINT_SETATTR, + OBD_CFG_REC = LLOG_OP_MAGIC | 0x20000, + /* PTL_CFG_REC = LLOG_OP_MAGIC | 0x30000, obsolete 1.4.0 */ + LLOG_GEN_REC = LLOG_OP_MAGIC | 0x40000, + /* LLOG_JOIN_REC = LLOG_OP_MAGIC | 0x50000, obsolete 1.8.0 */ + CHANGELOG_REC = LLOG_OP_MAGIC | 0x60000, + CHANGELOG_USER_REC = LLOG_OP_MAGIC | 0x70000, + LLOG_HDR_MAGIC = LLOG_OP_MAGIC | 0x45539, + LLOG_LOGID_MAGIC = LLOG_OP_MAGIC | 0x4553b, } llog_op_type; -/* - * for now, continue to support old pad records which have 0 for their - * type but still need to be swabbed for their length - */ -#define LLOG_REC_HDR_NEEDS_SWABBING(r) \ - (((r)->lrh_type & __swab32(LLOG_OP_MASK)) == \ - __swab32(LLOG_OP_MAGIC) || \ - (((r)->lrh_type == 0) && ((r)->lrh_len > LLOG_CHUNK_SIZE))) +#define LLOG_REC_HDR_NEEDS_SWABBING(r) \ + (((r)->lrh_type & __swab32(LLOG_OP_MASK)) == __swab32(LLOG_OP_MAGIC)) /** 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. */ struct llog_rec_hdr { - __u32 lrh_len; - __u32 lrh_index; - __u32 lrh_type; - __u32 lrh_padding; + __u32 lrh_len; + __u32 lrh_index; + __u32 lrh_type; + __u32 lrh_id; }; struct llog_rec_tail { - __u32 lrt_len; - __u32 lrt_index; + __u32 lrt_len; + __u32 lrt_index; }; struct llog_logid_rec { - struct llog_rec_hdr lid_hdr; - struct llog_logid lid_id; - __u32 lid_padding1; - __u32 lid_padding2; - __u32 lid_padding3; - __u32 lid_padding4; - __u32 lid_padding5; - struct llog_rec_tail lid_tail; -} __attribute__((packed)); - -struct llog_create_rec { - struct llog_rec_hdr lcr_hdr; - struct ll_fid lcr_fid; - obd_id lcr_oid; - obd_count lcr_oseq; - __u32 lcr_padding; - struct llog_rec_tail lcr_tail; -} __attribute__((packed)); - -struct llog_orphan_rec { - struct llog_rec_hdr lor_hdr; - obd_id lor_oid; - obd_count lor_ogen; - __u32 lor_padding; - struct llog_rec_tail lor_tail; + struct llog_rec_hdr lid_hdr; + struct llog_logid lid_id; + __u32 lid_padding1; + __u64 lid_padding2; + __u64 lid_padding3; + struct llog_rec_tail lid_tail; } __attribute__((packed)); struct llog_unlink_rec { - struct llog_rec_hdr lur_hdr; - obd_id lur_oid; - obd_count lur_oseq; - obd_count lur_count; - struct llog_rec_tail lur_tail; + struct llog_rec_hdr lur_hdr; + obd_id lur_oid; + obd_count lur_oseq; + obd_count lur_count; + struct llog_rec_tail lur_tail; } __attribute__((packed)); -struct llog_setattr_rec { - struct llog_rec_hdr lsr_hdr; - obd_id lsr_oid; - obd_count lsr_oseq; - __u32 lsr_uid; - __u32 lsr_gid; - __u32 lsr_padding; - struct llog_rec_tail lsr_tail; +struct llog_unlink64_rec { + struct llog_rec_hdr lur_hdr; + struct lu_fid lur_fid; + obd_count lur_count; /* to destroy the lost precreated */ + __u32 lur_padding1; + __u64 lur_padding2; + __u64 lur_padding3; + struct llog_rec_tail lur_tail; } __attribute__((packed)); struct llog_setattr64_rec { - struct llog_rec_hdr lsr_hdr; - obd_id lsr_oid; - obd_count lsr_oseq; - __u32 lsr_padding; - __u32 lsr_uid; - __u32 lsr_uid_h; - __u32 lsr_gid; - __u32 lsr_gid_h; - struct llog_rec_tail lsr_tail; + struct llog_rec_hdr lsr_hdr; + obd_id lsr_oid; + obd_seq lsr_oseq; + __u32 lsr_uid; + __u32 lsr_uid_h; + __u32 lsr_gid; + __u32 lsr_gid_h; + __u64 lsr_padding; + struct llog_rec_tail lsr_tail; } __attribute__((packed)); struct llog_size_change_rec { - struct llog_rec_hdr lsc_hdr; - struct ll_fid lsc_fid; - __u32 lsc_ioepoch; - __u32 lsc_padding; - struct llog_rec_tail lsc_tail; + struct llog_rec_hdr lsc_hdr; + struct ll_fid lsc_fid; + __u32 lsc_ioepoch; + __u32 lsc_padding1; + __u64 lsc_padding2; + __u64 lsc_padding3; + struct llog_rec_tail lsc_tail; } __attribute__((packed)); #define CHANGELOG_MAGIC 0xca103000 @@ -2690,16 +2841,21 @@ struct llog_changelog_user_rec { struct llog_rec_tail cur_tail; } __attribute__((packed)); +/* Old llog gen for compatibility */ struct llog_gen { - __u64 mnt_cnt; - __u64 conn_cnt; + __u64 mnt_cnt; + __u64 conn_cnt; } __attribute__((packed)); struct llog_gen_rec { - struct llog_rec_hdr lgr_hdr; - struct llog_gen lgr_gen; - struct llog_rec_tail lgr_tail; + struct llog_rec_hdr lgr_hdr; + struct llog_gen lgr_gen; + __u64 padding1; + __u64 padding2; + __u64 padding3; + struct llog_rec_tail lgr_tail; }; + /* On-disk header structure of each log object, stored in little endian order */ #define LLOG_CHUNK_SIZE 8192 #define LLOG_HEADER_SIZE (96) @@ -2708,9 +2864,11 @@ struct llog_gen_rec { #define LLOG_MIN_REC_SIZE (24) /* round(llog_rec_hdr + llog_rec_tail) */ /* flags for the logs */ -#define LLOG_F_ZAP_WHEN_EMPTY 0x1 -#define LLOG_F_IS_CAT 0x2 -#define LLOG_F_IS_PLAIN 0x4 +enum llog_flag { + LLOG_F_ZAP_WHEN_EMPTY = 0x1, + LLOG_F_IS_CAT = 0x2, + LLOG_F_IS_PLAIN = 0x4, +}; struct llog_log_hdr { struct llog_rec_hdr llh_hdr; @@ -2727,9 +2885,9 @@ struct llog_log_hdr { struct llog_rec_tail llh_tail; } __attribute__((packed)); -#define LLOG_BITMAP_SIZE(llh) ((llh->llh_hdr.lrh_len - \ - llh->llh_bitmap_offset - \ - sizeof(llh->llh_tail)) * 8) +#define LLOG_BITMAP_SIZE(llh) (__u32)((llh->llh_hdr.lrh_len - \ + llh->llh_bitmap_offset - \ + sizeof(llh->llh_tail)) * 8) /** log cookies are used to reference a specific log file and a record therein */ struct llog_cookie { @@ -2869,8 +3027,7 @@ extern void lustre_swab_lov_mds_md(struct lov_mds_md *lmm); extern void lustre_swab_llogd_body (struct llogd_body *d); extern void lustre_swab_llog_hdr (struct llog_log_hdr *h); extern void lustre_swab_llogd_conn_body (struct llogd_conn_body *d); -extern void lustre_swab_llog_rec(struct llog_rec_hdr *rec, - struct llog_rec_tail *tail); +extern void lustre_swab_llog_rec(struct llog_rec_hdr *rec); struct lustre_cfg; extern void lustre_swab_lustre_cfg(struct lustre_cfg *lcfg); @@ -2882,6 +3039,91 @@ void dump_obdo(struct obdo *oa); void dump_ost_body(struct ost_body *ob); void dump_rcs(__u32 *rc); +#define IDX_INFO_MAGIC 0x3D37CC37 + +/* Index file transfer through the network. The server serializes the index into + * a byte stream which is sent to the client via a bulk transfer */ +struct idx_info { + __u32 ii_magic; + + /* reply: see idx_info_flags below */ + __u32 ii_flags; + + /* request & reply: number of lu_idxpage (to be) transferred */ + __u16 ii_count; + __u16 ii_pad0; + + /* request: requested attributes passed down to the iterator API */ + __u32 ii_attrs; + + /* request & reply: index file identifier (FID) */ + struct lu_fid ii_fid; + + /* reply: version of the index file before starting to walk the index. + * Please note that the version can be modified at any time during the + * transfer */ + __u64 ii_version; + + /* request: hash to start with: + * reply: hash of the first entry of the first lu_idxpage and hash + * of the entry to read next if any */ + __u64 ii_hash_start; + __u64 ii_hash_end; + + /* reply: size of keys in lu_idxpages, minimal one if II_FL_VARKEY is + * set */ + __u16 ii_keysize; + + /* reply: size of records in lu_idxpages, minimal one if II_FL_VARREC + * is set */ + __u16 ii_recsize; + + __u32 ii_pad1; + __u64 ii_pad2; + __u64 ii_pad3; +}; +extern void lustre_swab_idx_info(struct idx_info *ii); + +#define II_END_OFF MDS_DIR_END_OFF /* all entries have been read */ + +/* List of flags used in idx_info::ii_flags */ +enum idx_info_flags { + II_FL_NOHASH = 1 << 0, /* client doesn't care about hash value */ + II_FL_VARKEY = 1 << 1, /* keys can be of variable size */ + II_FL_VARREC = 1 << 2, /* records can be of variable size */ + II_FL_NONUNQ = 1 << 3, /* index supports non-unique keys */ +}; + +#define LIP_MAGIC 0x8A6D6B6C + +/* 4KB (= LU_PAGE_SIZE) container gathering key/record pairs */ +struct lu_idxpage { + /* 16-byte header */ + __u32 lip_magic; + __u16 lip_flags; + __u16 lip_nr; /* number of entries in the container */ + __u64 lip_pad0; /* additional padding for future use */ + + /* key/record pairs are stored in the remaining 4080 bytes. + * depending upon the flags in idx_info::ii_flags, each key/record + * pair might be preceded by: + * - a hash value + * - the key size (II_FL_VARKEY is set) + * - the record size (II_FL_VARREC is set) + * + * For the time being, we only support fixed-size key & record. */ + char lip_entries[0]; +}; + +#define LIP_HDR_SIZE (offsetof(struct lu_idxpage, lip_entries)) + +/* Gather all possible type associated with a 4KB container */ +union lu_page { + struct lu_dirpage lp_dir; /* for MDS_READPAGE */ + struct lu_idxpage lp_idx; /* for OBD_IDX_READ */ + char lp_array[LU_PAGE_SIZE]; +}; + /* this will be used when OBD_CONNECT_CHANGE_QS is set */ struct qunit_data { /**