X-Git-Url: https://git.whamcloud.com/?p=fs%2Flustre-release.git;a=blobdiff_plain;f=lustre%2Finclude%2Flustre%2Flustre_idl.h;h=9330f950d0447bc38267732694effa941c127fd7;hp=a41e82a12dc3e1d48fec2976a2f389dc31456a53;hb=79d06b6fac3d18f2e5755a940e8afa42e70f3ba2;hpb=fa32b478599bbdfe6d51e41d8715665632d011b3;ds=sidebyside diff --git a/lustre/include/lustre/lustre_idl.h b/lustre/include/lustre/lustre_idl.h index a41e82a..9330f95 100644 --- a/lustre/include/lustre/lustre_idl.h +++ b/lustre/include/lustre/lustre_idl.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) 2011, 2012, Whamcloud, Inc. + * Copyright (c) 2011, 2012, Intel Corporation. */ /* * This file is part of Lustre, http://www.lustre.org/ @@ -265,6 +265,8 @@ static inline int range_compare_loc(const struct lu_seq_range *r1, /** * Flags for lustre_mdt_attrs::lma_compat and lustre_mdt_attrs::lma_incompat. + * Deprecated since HSM and SOM attributes are now stored in separate on-disk + * xattr. */ enum lma_compat { LMAC_HSM = 0x00000001, @@ -275,9 +277,10 @@ enum lma_compat { * Masks for all features that should be supported by a Lustre version to * access a specific file. * This information is stored in lustre_mdt_attrs::lma_incompat. - * - * NOTE: No incompat feature should be added before bug #17670 is landed. */ +enum lma_incompat { + LMAI_RELEASED = 0x0000001, /* file is released */ +}; #define LMA_INCOMPAT_SUPP 0x0 /** @@ -300,66 +303,56 @@ struct lustre_mdt_attrs { struct lu_fid lma_self_fid; /** mdt/ost type, others */ __u64 lma_flags; - /* IO Epoch SOM attributes belongs to */ - __u64 lma_ioepoch; - /** total file size in objects */ - __u64 lma_som_size; - /** total fs blocks in objects */ - __u64 lma_som_blocks; - /** mds mount id the size is valid for */ - __u64 lma_som_mountid; }; /** - * Fill \a lma with its first content. - * Only fid is stored. + * Prior to 2.4, the LMA structure also included SOM attributes which has since + * been moved to a dedicated xattr */ -static inline void lustre_lma_init(struct lustre_mdt_attrs *lma, - const struct lu_fid *fid) -{ - lma->lma_compat = 0; - lma->lma_incompat = 0; - memcpy(&lma->lma_self_fid, fid, sizeof(*fid)); - lma->lma_flags = 0; - lma->lma_ioepoch = 0; - lma->lma_som_size = 0; - lma->lma_som_blocks = 0; - lma->lma_som_mountid = 0; +#define LMA_OLD_SIZE (sizeof(struct lustre_mdt_attrs) + 4 * sizeof(__u64)) + +extern void lustre_lma_swab(struct lustre_mdt_attrs *lma); +extern void lustre_lma_init(struct lustre_mdt_attrs *lma, + const struct lu_fid *fid); +/** + * SOM on-disk attributes stored in a separate xattr. + */ +struct som_attrs { + /** Bitfield for supported data in this structure. For future use. */ + __u32 som_compat; + + /** Incompat feature list. The supported feature mask is availabe in + * SOM_INCOMPAT_SUPP */ + __u32 som_incompat; - /* If a field is added in struct lustre_mdt_attrs, zero it explicitly - * and change the test below. */ - LASSERT(sizeof(*lma) == - (offsetof(struct lustre_mdt_attrs, lma_som_mountid) + - sizeof(lma->lma_som_mountid))); + /** IO Epoch SOM attributes belongs to */ + __u64 som_ioepoch; + /** total file size in objects */ + __u64 som_size; + /** total fs blocks in objects */ + __u64 som_blocks; + /** mds mount id the size is valid for */ + __u64 som_mountid; }; +extern void lustre_som_swab(struct som_attrs *attrs); -extern void lustre_swab_lu_fid(struct lu_fid *fid); +#define SOM_INCOMPAT_SUPP 0x0 /** - * Swab, if needed, lustre_mdt_attr struct to on-disk format. - * Otherwise, do not touch it. + * HSM on-disk attributes stored in a separate xattr. */ -static inline void lustre_lma_swab(struct lustre_mdt_attrs *lma) -{ - /* Use LUSTRE_MSG_MAGIC to detect local endianess. */ - if (LUSTRE_MSG_MAGIC != cpu_to_le32(LUSTRE_MSG_MAGIC)) { - __swab32s(&lma->lma_compat); - __swab32s(&lma->lma_incompat); - lustre_swab_lu_fid(&lma->lma_self_fid); - __swab64s(&lma->lma_flags); - __swab64s(&lma->lma_ioepoch); - __swab64s(&lma->lma_som_size); - __swab64s(&lma->lma_som_blocks); - __swab64s(&lma->lma_som_mountid); - } -}; - -/* This is the maximum number of MDTs allowed in CMD testing until such - * a time that FID-on-OST is implemented. This is due to the limitations - * of packing non-0-MDT numbers into the FID SEQ namespace. Once FID-on-OST - * is implemented this limit will be virtually unlimited. */ -#define MAX_MDT_COUNT 8 +struct hsm_attrs { + /** Bitfield for supported data in this structure. For future use. */ + __u32 hsm_compat; + /** HSM flags, see hsm_flags enum below */ + __u32 hsm_flags; + /** backend archive id associated with the file */ + __u64 hsm_arch_id; + /** version associated with the last archiving, if any */ + __u64 hsm_arch_ver; +}; +extern void lustre_hsm_swab(struct hsm_attrs *attrs); /** * fid constants @@ -406,30 +399,34 @@ static inline obd_id fid_ver_oid(const struct lu_fid *fid) * http://arch.lustre.org/index.php?title=Interoperability_fids_zfs#NEW.0 */ enum fid_seq { - FID_SEQ_OST_MDT0 = 0, - FID_SEQ_LLOG = 1, - FID_SEQ_ECHO = 2, - FID_SEQ_OST_MDT1 = 3, - FID_SEQ_OST_MAX = 9, /* Max MDT count before OST_on_FID */ - FID_SEQ_RSVD = 11, - FID_SEQ_IGIF = 12, - FID_SEQ_IGIF_MAX = 0x0ffffffffULL, - FID_SEQ_IDIF = 0x100000000ULL, - FID_SEQ_IDIF_MAX = 0x1ffffffffULL, - /* Normal FID sequence starts from this value, i.e. 1<<33 */ - FID_SEQ_START = 0x200000000ULL, + FID_SEQ_OST_MDT0 = 0, + FID_SEQ_LLOG = 1, + FID_SEQ_ECHO = 2, + FID_SEQ_OST_MDT1 = 3, + FID_SEQ_OST_MAX = 9, /* Max MDT count before OST_on_FID */ + FID_SEQ_RSVD = 11, + FID_SEQ_IGIF = 12, + FID_SEQ_IGIF_MAX = 0x0ffffffffULL, + FID_SEQ_IDIF = 0x100000000ULL, + 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 */ + FID_SEQ_LOCAL_FILE = 0x200000001ULL, + FID_SEQ_DOT_LUSTRE = 0x200000002ULL, /* sequence is used for local named objects FIDs generated * by local_object_storage library */ - FID_SEQ_SPECIAL = 0x200000004ULL, - FID_SEQ_QUOTA = 0x200000005ULL, - FID_SEQ_QUOTA_GLB = 0x200000006ULL, - FID_SEQ_LOCAL_NAME = 0x200000007ULL, - FID_SEQ_NORMAL = 0x200000400ULL, - FID_SEQ_LOV_DEFAULT= 0xffffffffffffffffULL + FID_SEQ_LOCAL_NAME = 0x200000003ULL, + /* Because current FLD will only cache the fid sequence, instead + * of oid on the client side, if the FID needs to be exposed to + * clients sides, it needs to make sure all of fids under one + * sequence will be located in one MDT. */ + FID_SEQ_SPECIAL = 0x200000004ULL, + FID_SEQ_QUOTA = 0x200000005ULL, + FID_SEQ_QUOTA_GLB = 0x200000006ULL, + FID_SEQ_ROOT = 0x200000007ULL, /* Located on MDT0 */ + FID_SEQ_NORMAL = 0x200000400ULL, + FID_SEQ_LOV_DEFAULT = 0xffffffffffffffffULL }; #define OBIF_OID_MAX_BITS 32 @@ -456,27 +453,58 @@ static inline int fid_seq_is_mdt0(obd_seq seq) return (seq == FID_SEQ_OST_MDT0); } -static inline int fid_seq_is_cmd(const __u64 seq) +static inline int fid_seq_is_mdt(const __u64 seq) { - return (seq >= FID_SEQ_OST_MDT1 && seq <= FID_SEQ_OST_MAX); + return seq == FID_SEQ_OST_MDT0 || seq >= FID_SEQ_NORMAL; }; -static inline int fid_seq_is_mdt(const __u64 seq) +static inline int fid_seq_is_echo(obd_seq seq) { - return seq == FID_SEQ_OST_MDT0 || - (seq >= FID_SEQ_OST_MDT1 && seq <= FID_SEQ_OST_MAX); -}; + return (seq == FID_SEQ_ECHO); +} + +static inline int fid_is_echo(const struct lu_fid *fid) +{ + return fid_seq_is_echo(fid_seq(fid)); +} + +static inline int fid_seq_is_llog(obd_seq seq) +{ + return (seq == FID_SEQ_LLOG); +} + +static inline int fid_is_llog(const struct lu_fid *fid) +{ + return fid_seq_is_llog(fid_seq(fid)); +} static inline int fid_seq_is_rsvd(const __u64 seq) { return (seq > FID_SEQ_OST_MDT0 && seq <= FID_SEQ_RSVD); }; +static inline int fid_seq_is_special(const __u64 seq) +{ + return seq == FID_SEQ_SPECIAL; +}; + +static inline int fid_seq_is_local_file(const __u64 seq) +{ + return seq == FID_SEQ_LOCAL_FILE; +}; + static inline int fid_is_mdt0(const struct lu_fid *fid) { return fid_seq_is_mdt0(fid_seq(fid)); } +static inline void lu_root_fid(struct lu_fid *fid) +{ + fid->f_seq = FID_SEQ_ROOT; + fid->f_oid = 1; + fid->f_ver = 0; +} + /** * Check if a fid is igif or not. * \param fid the fid to be tested. @@ -512,6 +540,11 @@ struct ost_id { obd_seq oi_seq; }; +static inline int fid_is_local_file(const struct lu_fid *fid) +{ + return fid_seq_is_local_file(fid_seq(fid)); +} + static inline int fid_seq_is_norm(const __u64 seq) { return (seq >= FID_SEQ_NORMAL); @@ -534,6 +567,13 @@ 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) @@ -673,6 +713,13 @@ static inline obd_id ostid_id(struct ost_id *ostid) return ostid->oi_id; } +/* Check whether the fid is for LAST_ID */ +static inline int fid_is_last_id(const struct lu_fid *fid) +{ + return (fid_is_idif(fid) || fid_is_norm(fid) || fid_is_echo(fid)) && + fid_oid(fid) == 0; +} + /** * Get inode number from a igif. * \param fid a igif to get inode number from. @@ -684,21 +731,6 @@ static inline ino_t lu_igif_ino(const struct lu_fid *fid) } /** - * Build igif from the inode number/generation. - */ -#define LU_IGIF_BUILD(fid, ino, gen) \ -do { \ - fid->f_seq = ino; \ - fid->f_oid = gen; \ - fid->f_ver = 0; \ -} while(0) -static inline void lu_igif_build(struct lu_fid *fid, __u32 ino, __u32 gen) -{ - LU_IGIF_BUILD(fid, ino, gen); - LASSERT(fid_is_igif(fid)); -} - -/** * Get inode generation from a igif. * \param fid a igif to get inode generation from. * \return inode generation for the igif. @@ -708,65 +740,70 @@ static inline __u32 lu_igif_gen(const struct lu_fid *fid) return fid_oid(fid); } +/** + * Build igif from the inode number/generation. + */ +static inline void lu_igif_build(struct lu_fid *fid, __u32 ino, __u32 gen) +{ + fid->f_seq = ino; + fid->f_oid = gen; + fid->f_ver = 0; +} + /* * Fids are transmitted across network (in the sender byte-ordering), * and stored on disk in big-endian order. */ static inline void fid_cpu_to_le(struct lu_fid *dst, const struct lu_fid *src) { - /* check that all fields are converted */ - CLASSERT(sizeof *src == - sizeof fid_seq(src) + - sizeof fid_oid(src) + sizeof fid_ver(src)); - LASSERTF(fid_is_igif(src) || fid_ver(src) == 0, DFID"\n", PFID(src)); - dst->f_seq = cpu_to_le64(fid_seq(src)); - dst->f_oid = cpu_to_le32(fid_oid(src)); - dst->f_ver = cpu_to_le32(fid_ver(src)); + /* check that all fields are converted */ + CLASSERT(sizeof *src == + sizeof fid_seq(src) + + sizeof fid_oid(src) + sizeof fid_ver(src)); + dst->f_seq = cpu_to_le64(fid_seq(src)); + dst->f_oid = cpu_to_le32(fid_oid(src)); + dst->f_ver = cpu_to_le32(fid_ver(src)); } static inline void fid_le_to_cpu(struct lu_fid *dst, const struct lu_fid *src) { - /* check that all fields are converted */ - CLASSERT(sizeof *src == - sizeof fid_seq(src) + - sizeof fid_oid(src) + sizeof fid_ver(src)); - dst->f_seq = le64_to_cpu(fid_seq(src)); - dst->f_oid = le32_to_cpu(fid_oid(src)); - dst->f_ver = le32_to_cpu(fid_ver(src)); - LASSERTF(fid_is_igif(dst) || fid_ver(dst) == 0, DFID"\n", PFID(dst)); + /* check that all fields are converted */ + CLASSERT(sizeof *src == + sizeof fid_seq(src) + + sizeof fid_oid(src) + sizeof fid_ver(src)); + dst->f_seq = le64_to_cpu(fid_seq(src)); + dst->f_oid = le32_to_cpu(fid_oid(src)); + dst->f_ver = le32_to_cpu(fid_ver(src)); } static inline void fid_cpu_to_be(struct lu_fid *dst, const struct lu_fid *src) { - /* check that all fields are converted */ - CLASSERT(sizeof *src == - sizeof fid_seq(src) + - sizeof fid_oid(src) + sizeof fid_ver(src)); - LASSERTF(fid_is_igif(src) || fid_ver(src) == 0, DFID"\n", PFID(src)); - dst->f_seq = cpu_to_be64(fid_seq(src)); - dst->f_oid = cpu_to_be32(fid_oid(src)); - dst->f_ver = cpu_to_be32(fid_ver(src)); + /* check that all fields are converted */ + CLASSERT(sizeof *src == + sizeof fid_seq(src) + + sizeof fid_oid(src) + sizeof fid_ver(src)); + dst->f_seq = cpu_to_be64(fid_seq(src)); + dst->f_oid = cpu_to_be32(fid_oid(src)); + dst->f_ver = cpu_to_be32(fid_ver(src)); } static inline void fid_be_to_cpu(struct lu_fid *dst, const struct lu_fid *src) { - /* check that all fields are converted */ - CLASSERT(sizeof *src == - sizeof fid_seq(src) + - sizeof fid_oid(src) + sizeof fid_ver(src)); - dst->f_seq = be64_to_cpu(fid_seq(src)); - dst->f_oid = be32_to_cpu(fid_oid(src)); - dst->f_ver = be32_to_cpu(fid_ver(src)); - LASSERTF(fid_is_igif(dst) || fid_ver(dst) == 0, DFID"\n", PFID(dst)); + /* check that all fields are converted */ + CLASSERT(sizeof *src == + sizeof fid_seq(src) + + sizeof fid_oid(src) + sizeof fid_ver(src)); + dst->f_seq = be64_to_cpu(fid_seq(src)); + dst->f_oid = be32_to_cpu(fid_oid(src)); + dst->f_ver = be32_to_cpu(fid_ver(src)); } static inline int fid_is_sane(const struct lu_fid *fid) { - return - fid != NULL && - ((fid_seq(fid) >= FID_SEQ_START && fid_oid(fid) != 0 - && fid_ver(fid) == 0) || - fid_is_igif(fid) || fid_seq_is_rsvd(fid_seq(fid))); + return fid != NULL && + ((fid_seq(fid) >= FID_SEQ_START && fid_ver(fid) == 0) || + fid_is_igif(fid) || fid_is_idif(fid) || + fid_seq_is_rsvd(fid_seq(fid))); } static inline int fid_is_zero(const struct lu_fid *fid) @@ -777,17 +814,12 @@ static inline int fid_is_zero(const struct lu_fid *fid) extern void lustre_swab_lu_fid(struct lu_fid *fid); extern void lustre_swab_lu_seq_range(struct lu_seq_range *range); -static inline int lu_fid_eq(const struct lu_fid *f0, - const struct lu_fid *f1) +static inline int lu_fid_eq(const struct lu_fid *f0, const struct lu_fid *f1) { - /* Check that there is no alignment padding. */ - CLASSERT(sizeof *f0 == - sizeof f0->f_seq + sizeof f0->f_oid + sizeof f0->f_ver); - LASSERTF((fid_is_igif(f0) || fid_is_idif(f0)) || - fid_ver(f0) == 0, DFID, PFID(f0)); - LASSERTF((fid_is_igif(f1) || fid_is_idif(f1)) || - fid_ver(f1) == 0, DFID, PFID(f1)); - return memcmp(f0, f1, sizeof *f0) == 0; + /* Check that there is no alignment padding. */ + CLASSERT(sizeof *f0 == + sizeof f0->f_seq + sizeof f0->f_oid + sizeof f0->f_ver); + return memcmp(f0, f1, sizeof *f0) == 0; } #define __diff_normalize(val0, val1) \ @@ -819,11 +851,27 @@ static inline int lu_fid_cmp(const struct lu_fid *f0, * enumeration. */ enum lu_dirent_attrs { - LUDA_FID = 0x0001, - LUDA_TYPE = 0x0002, - LUDA_64BITHASH = 0x0004, + LUDA_FID = 0x0001, + LUDA_TYPE = 0x0002, + LUDA_64BITHASH = 0x0004, + + /* The following attrs are used for MDT interanl only, + * not visible to client */ + + /* Verify the dirent consistency */ + LUDA_VERIFY = 0x8000, + /* Only check but not repair the dirent inconsistency */ + LUDA_VERIFY_DRYRUN = 0x4000, + /* The dirent has been repaired, or to be repaired (dryrun). */ + LUDA_REPAIR = 0x2000, + /* The system is upgraded, has beed or to be repaired (dryrun). */ + LUDA_UPGRADE = 0x1000, + /* Ignore this record, go to next directly. */ + LUDA_IGNORE = 0x0800, }; +#define LU_DIRENT_ATTRS_MASK 0xf800 + /** * Layout of readdir pages, as transmitted on wire. */ @@ -1130,7 +1178,7 @@ extern void lustre_swab_ptlrpc_body(struct ptlrpc_body *pb); #define OBD_CONNECT_RMT_CLIENT 0x10000ULL /*Remote client */ #define OBD_CONNECT_RMT_CLIENT_FORCE 0x20000ULL /*Remote client by force */ #define OBD_CONNECT_BRW_SIZE 0x40000ULL /*Max bytes per rpc */ -#define OBD_CONNECT_QUOTA64 0x80000ULL /*64bit qunit_data.qd_count */ +#define OBD_CONNECT_QUOTA64 0x80000ULL /*Not used since 2.4 */ #define OBD_CONNECT_MDS_CAPA 0x100000ULL /*MDS capability */ #define OBD_CONNECT_OSS_CAPA 0x200000ULL /*OSS capability */ #define OBD_CONNECT_CANCELSET 0x400000ULL /*Early batched cancels. */ @@ -1139,7 +1187,7 @@ extern void lustre_swab_ptlrpc_body(struct ptlrpc_body *pb); #define OBD_CONNECT_LRU_RESIZE 0x2000000ULL /*LRU resize feature. */ #define OBD_CONNECT_MDS_MDS 0x4000000ULL /*MDS-MDS connection */ #define OBD_CONNECT_REAL 0x8000000ULL /*real connection */ -#define OBD_CONNECT_CHANGE_QS 0x10000000ULL /*shrink/enlarge qunit */ +#define OBD_CONNECT_CHANGE_QS 0x10000000ULL /*Not used since 2.4 */ #define OBD_CONNECT_CKSUM 0x20000000ULL /*support several cksum algos*/ #define OBD_CONNECT_FID 0x40000000ULL /*FID is supported by server */ #define OBD_CONNECT_VBR 0x80000000ULL /*version based recovery */ @@ -1159,9 +1207,13 @@ extern void lustre_swab_ptlrpc_body(struct ptlrpc_body *pb); * 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_FLOCK_OWNER 0x200000000000ULL /* for the fixed 1.8 + * policy and 2.x server */ #define OBD_CONNECT_LVB_TYPE 0x400000000000ULL /* variable type of LVB */ +#define OBD_CONNECT_NANOSEC_TIME 0x800000000000ULL /* nanosecond timestamps */ #define OBD_CONNECT_LIGHTWEIGHT 0x1000000000000ULL/* lightweight connection */ +#define OBD_CONNECT_SHORTIO 0x2000000000000ULL/* short io */ +#define OBD_CONNECT_PINGLESS 0x4000000000000ULL/* pings not required */ /* 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 @@ -1200,15 +1252,15 @@ extern void lustre_swab_ptlrpc_body(struct ptlrpc_body *pb); OBD_CONNECT_SOM | OBD_CONNECT_FULL20 | \ OBD_CONNECT_64BITHASH | OBD_CONNECT_JOBSTATS | \ OBD_CONNECT_EINPROGRESS | \ - OBD_CONNECT_LIGHTWEIGHT) + OBD_CONNECT_LIGHTWEIGHT | OBD_CONNECT_UMASK | \ + OBD_CONNECT_LVB_TYPE | OBD_CONNECT_LAYOUTLOCK |\ + OBD_CONNECT_PINGLESS) #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_BRW_SIZE | OBD_CONNECT_OSS_CAPA | \ OBD_CONNECT_CANCELSET | OBD_CONNECT_AT | \ LRU_RESIZE_CONNECT_FLAG | OBD_CONNECT_CKSUM | \ - OBD_CONNECT_CHANGE_QS | \ - OBD_CONNECT_OSS_CAPA | \ OBD_CONNECT_RMT_CLIENT | \ OBD_CONNECT_RMT_CLIENT_FORCE | OBD_CONNECT_VBR | \ OBD_CONNECT_MDS | OBD_CONNECT_SKIP_ORPHAN | \ @@ -1216,11 +1268,14 @@ extern void lustre_swab_ptlrpc_body(struct ptlrpc_body *pb); OBD_CONNECT_64BITHASH | OBD_CONNECT_MAXBYTES | \ OBD_CONNECT_MAX_EASIZE | \ OBD_CONNECT_EINPROGRESS | \ - OBD_CONNECT_JOBSTATS | OBD_CONNECT_LIGHTWEIGHT) + OBD_CONNECT_JOBSTATS | \ + OBD_CONNECT_LIGHTWEIGHT | OBD_CONNECT_LVB_TYPE|\ + OBD_CONNECT_LAYOUTLOCK | OBD_CONNECT_FID | \ + OBD_CONNECT_PINGLESS) #define ECHO_CONNECT_SUPPORTED (0) #define MGS_CONNECT_SUPPORTED (OBD_CONNECT_VERSION | OBD_CONNECT_AT | \ OBD_CONNECT_FULL20 | OBD_CONNECT_IMP_RECOV | \ - OBD_CONNECT_MNE_SWAB) + OBD_CONNECT_MNE_SWAB | OBD_CONNECT_PINGLESS) /* Features required for this version of the client to work with server */ #define CLIENT_CONNECT_MDT_REQD (OBD_CONNECT_IBITS | OBD_CONNECT_FID | \ @@ -1340,7 +1395,7 @@ typedef enum { OST_SET_INFO = 17, OST_QUOTACHECK = 18, OST_QUOTACTL = 19, - OST_QUOTA_ADJUST_QUNIT = 20, + OST_QUOTA_ADJUST_QUNIT = 20, /* not used since 2.4 */ OST_LAST_OPC } ost_cmd_t; #define OST_FIRST_OPC OST_REPLY @@ -1383,6 +1438,21 @@ 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 */ @@ -1427,6 +1497,9 @@ struct lov_mds_md_v1 { /* LOV EA mds/wire data (little-endian) */ #define XATTR_NAME_LINK "trusted.link" #define XATTR_NAME_FID "trusted.fid" #define XATTR_NAME_VERSION "trusted.version" +#define XATTR_NAME_SOM "trusted.som" +#define XATTR_NAME_HSM "trusted.hsm" +#define XATTR_NAME_LFSCK_NAMESPACE "trusted.lfsck_namespace" struct lov_mds_md_v3 { /* LOV EA mds/wire data (little-endian) */ @@ -1479,7 +1552,10 @@ struct lov_mds_md_v3 { /* LOV EA mds/wire data (little-endian) */ #define OBD_MD_MDS (0x0000000100000000ULL) /* where an inode lives on */ #define OBD_MD_REINT (0x0000000200000000ULL) /* reintegrate oa */ #define OBD_MD_MEA (0x0000000400000000ULL) /* CMD split EA */ -#define OBD_MD_MDTIDX (0x0000000800000000ULL) /* Get MDT index */ + +/* OBD_MD_MDTIDX is used to get MDT index, but it is never been used overwire, + * and it is already obsolete since 2.3 */ +/* #define OBD_MD_MDTIDX (0x0000000800000000ULL) */ #define OBD_MD_FLXATTR (0x0000001000000000ULL) /* xattr */ #define OBD_MD_FLXATTRLS (0x0000002000000000ULL) /* xattr list */ @@ -1510,6 +1586,21 @@ struct lov_mds_md_v3 { /* LOV EA mds/wire data (little-endian) */ /* don't forget obdo_fid which is way down at the bottom so it can * come after the definition of llog_cookie */ +enum hss_valid { + HSS_SETMASK = 0x01, + HSS_CLEARMASK = 0x02, + HSS_ARCHIVE_ID = 0x04, +}; + +struct hsm_state_set { + __u32 hss_valid; + __u32 hss_archive_id; + __u64 hss_setmask; + __u64 hss_clearmask; +}; + +extern void lustre_swab_hsm_user_state(struct hsm_user_state *hus); +extern void lustre_swab_hsm_state_set(struct hsm_state_set *hss); extern void lustre_swab_obd_statfs (struct obd_statfs *os); @@ -1538,12 +1629,14 @@ extern void lustre_swab_obd_statfs (struct obd_statfs *os); #define OST_MAX_PRECREATE 20000 struct obd_ioobj { - obd_id ioo_id; - obd_seq ioo_seq; - __u32 ioo_type; - __u32 ioo_bufcnt; + struct ost_id ioo_oid; + __u32 ioo_type; + __u32 ioo_bufcnt; }; +#define ioo_id ioo_oid.oi_id +#define ioo_seq ioo_oid.oi_seq + extern void lustre_swab_obd_ioobj (struct obd_ioobj *ioo); /* multiple of 8 bytes => can array */ @@ -1567,59 +1660,271 @@ extern void lustre_swab_niobuf_remote (struct niobuf_remote *nbr); do { blocks = OST_LVB_ERR_INIT + rc; } while (0) #define OST_LVB_GET_ERR(blocks) (int)(blocks - OST_LVB_ERR_INIT) +struct ost_lvb_v1 { + __u64 lvb_size; + obd_time lvb_mtime; + obd_time lvb_atime; + obd_time lvb_ctime; + __u64 lvb_blocks; +}; + +extern void lustre_swab_ost_lvb_v1(struct ost_lvb_v1 *lvb); + struct ost_lvb { - __u64 lvb_size; - obd_time lvb_mtime; - obd_time lvb_atime; - obd_time lvb_ctime; - __u64 lvb_blocks; + __u64 lvb_size; + obd_time lvb_mtime; + obd_time lvb_atime; + obd_time lvb_ctime; + __u64 lvb_blocks; + __u32 lvb_mtime_ns; + __u32 lvb_atime_ns; + __u32 lvb_ctime_ns; + __u32 lvb_padding; +}; + +extern void lustre_swab_ost_lvb(struct ost_lvb *lvb); + +/* + * lquota data structures + */ + +#ifndef QUOTABLOCK_BITS +#define QUOTABLOCK_BITS 10 +#endif + +#ifndef QUOTABLOCK_SIZE +#define QUOTABLOCK_SIZE (1 << QUOTABLOCK_BITS) +#endif + +#ifndef toqb +#define toqb(x) (((x) + QUOTABLOCK_SIZE - 1) >> QUOTABLOCK_BITS) +#endif + +/* 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 */ +}; + +/* quotactl management */ +struct obd_quotactl { + __u32 qc_cmd; + __u32 qc_type; /* see Q_* flag below */ + __u32 qc_id; + __u32 qc_stat; + struct obd_dqinfo qc_dqinfo; + struct obd_dqblk qc_dqblk; +}; + +extern void lustre_swab_obd_quotactl(struct obd_quotactl *q); + +#define Q_QUOTACHECK 0x800100 /* deprecated as of 2.4 */ +#define Q_INITQUOTA 0x800101 /* deprecated as of 2.4 */ +#define Q_GETOINFO 0x800102 /* get obd quota info */ +#define Q_GETOQUOTA 0x800103 /* get obd quotas */ +#define Q_FINVALIDATE 0x800104 /* deprecated as of 2.4 */ + +#define Q_COPY(out, in, member) (out)->member = (in)->member + +#define QCTL_COPY(out, in) \ +do { \ + Q_COPY(out, in, qc_cmd); \ + Q_COPY(out, in, qc_type); \ + Q_COPY(out, in, qc_id); \ + Q_COPY(out, in, qc_stat); \ + Q_COPY(out, in, qc_dqinfo); \ + Q_COPY(out, in, qc_dqblk); \ +} while (0) + +/* Body of quota request used for quota acquire/release RPCs between quota + * master (aka QMT) and slaves (ak QSD). */ +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 below */ + __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 + +#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 */ + +extern void lustre_swab_quota_body(struct quota_body *b); + +/* 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 */ +}; + +/* 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_time; + __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; +}; + +extern void lustre_swab_lquota_lvb(struct lquota_lvb *lvb); + +/* LVB used with global quota lock */ +#define lvb_glb_ver lvb_id_may_rel /* current version of the global index */ + +/* op codes */ +typedef enum { + QUOTA_DQACQ = 601, + QUOTA_DQREL = 602, + QUOTA_LAST_OPC +} quota_cmd_t; +#define QUOTA_FIRST_OPC QUOTA_DQACQ + +/* * MDS REQ RECORDS */ /* opcodes */ typedef enum { - MDS_GETATTR = 33, - MDS_GETATTR_NAME = 34, - MDS_CLOSE = 35, - MDS_REINT = 36, - MDS_READPAGE = 37, - MDS_CONNECT = 38, - MDS_DISCONNECT = 39, - MDS_GETSTATUS = 40, - MDS_STATFS = 41, - MDS_PIN = 42, - MDS_UNPIN = 43, - MDS_SYNC = 44, - MDS_DONE_WRITING = 45, - MDS_SET_INFO = 46, - MDS_QUOTACHECK = 47, - MDS_QUOTACTL = 48, - MDS_GETXATTR = 49, - MDS_SETXATTR = 50, /* obsolete, now it's MDS_REINT op */ - MDS_WRITEPAGE = 51, - MDS_IS_SUBDIR = 52, - MDS_GET_INFO = 53, - MDS_LAST_OPC + MDS_GETATTR = 33, + MDS_GETATTR_NAME = 34, + MDS_CLOSE = 35, + MDS_REINT = 36, + MDS_READPAGE = 37, + MDS_CONNECT = 38, + MDS_DISCONNECT = 39, + MDS_GETSTATUS = 40, + MDS_STATFS = 41, + MDS_PIN = 42, + MDS_UNPIN = 43, + MDS_SYNC = 44, + MDS_DONE_WRITING = 45, + MDS_SET_INFO = 46, + MDS_QUOTACHECK = 47, + MDS_QUOTACTL = 48, + MDS_GETXATTR = 49, + MDS_SETXATTR = 50, /* obsolete, now it's MDS_REINT op */ + MDS_WRITEPAGE = 51, + MDS_IS_SUBDIR = 52, + MDS_GET_INFO = 53, + MDS_HSM_STATE_GET = 54, + MDS_HSM_STATE_SET = 55, + MDS_HSM_ACTION = 56, + MDS_HSM_PROGRESS = 57, + MDS_HSM_REQUEST = 58, + MDS_HSM_CT_REGISTER = 59, + MDS_HSM_CT_UNREGISTER = 60, + MDS_SWAP_LAYOUTS = 61, + MDS_LAST_OPC } mds_cmd_t; #define MDS_FIRST_OPC MDS_GETATTR + +/* opcodes for object update */ +typedef enum { + UPDATE_OBJ = 1000, + UPDATE_LAST_OPC +} update_cmd_t; + +#define UPDATE_FIRST_OPC UPDATE_OBJ + /* * Do not exceed 63 */ typedef enum { - REINT_SETATTR = 1, - REINT_CREATE = 2, - REINT_LINK = 3, - REINT_UNLINK = 4, - REINT_RENAME = 5, - REINT_OPEN = 6, - REINT_SETXATTR = 7, -// REINT_CLOSE = 8, + REINT_SETATTR = 1, + REINT_CREATE = 2, + REINT_LINK = 3, + REINT_UNLINK = 4, + REINT_RENAME = 5, + REINT_OPEN = 6, + REINT_SETXATTR = 7, + REINT_RMENTRY = 8, // REINT_WRITE = 9, REINT_MAX } mds_reint_t, mdt_reint_t; @@ -1643,11 +1948,9 @@ extern void lustre_swab_generic_32s (__u32 *val); #define MDS_INODELOCK_UPDATE 0x000002 /* size, links, timestamps */ #define MDS_INODELOCK_OPEN 0x000004 /* For opened files */ #define MDS_INODELOCK_LAYOUT 0x000008 /* for layout */ +#define MDS_INODELOCK_PERM 0x000010 /* for permission */ -/* Do not forget to increase MDS_INODELOCK_MAXSHIFT when adding new bits - * XXX: MDS_INODELOCK_MAXSHIFT should be increased to 3 once the layout lock is - * supported */ -#define MDS_INODELOCK_MAXSHIFT 2 +#define MDS_INODELOCK_MAXSHIFT 4 /* This FULL lock is useful to take on unlink sort of operations */ #define MDS_INODELOCK_FULL ((1<<(MDS_INODELOCK_MAXSHIFT+1))-1) @@ -1660,6 +1963,8 @@ 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 }; @@ -1668,19 +1973,20 @@ enum { /* mdt_thread_info.mti_flags. */ enum md_op_flags { - /* The flag indicates Size-on-MDS attributes are changed. */ - MF_SOM_CHANGE = (1 << 0), - /* Flags indicates an epoch opens or closes. */ - MF_EPOCH_OPEN = (1 << 1), - MF_EPOCH_CLOSE = (1 << 2), - MF_MDC_CANCEL_FID1 = (1 << 3), - MF_MDC_CANCEL_FID2 = (1 << 4), - MF_MDC_CANCEL_FID3 = (1 << 5), - MF_MDC_CANCEL_FID4 = (1 << 6), - /* There is a pending attribute update. */ - MF_SOM_AU = (1 << 7), - /* Cancel OST locks while getattr OST attributes. */ - MF_GETATTR_LOCK = (1 << 8), + /* The flag indicates Size-on-MDS attributes are changed. */ + MF_SOM_CHANGE = (1 << 0), + /* Flags indicates an epoch opens or closes. */ + MF_EPOCH_OPEN = (1 << 1), + MF_EPOCH_CLOSE = (1 << 2), + MF_MDC_CANCEL_FID1 = (1 << 3), + MF_MDC_CANCEL_FID2 = (1 << 4), + MF_MDC_CANCEL_FID3 = (1 << 5), + MF_MDC_CANCEL_FID4 = (1 << 6), + /* There is a pending attribute update. */ + MF_SOM_AU = (1 << 7), + /* Cancel OST locks while getattr OST attributes. */ + MF_GETATTR_LOCK = (1 << 8), + MF_GET_MDT_IDX = (1 << 9), }; #define MF_SOM_LOCAL_FLAGS (MF_SOM_CHANGE | MF_EPOCH_OPEN | MF_EPOCH_CLOSE) @@ -1736,7 +2042,7 @@ struct mdt_body { obd_time ctime; __u64 blocks; /* XID, in the case of MDS_READPAGE */ __u64 ioepoch; - __u64 ino; + __u64 unused1; /* was "ino" until 2.4.0 */ __u32 fsuid; __u32 fsgid; __u32 capability; @@ -1746,7 +2052,7 @@ struct mdt_body { __u32 flags; /* from vfs for pin/unpin, LUSTRE_BFLAG close */ __u32 rdev; __u32 nlink; /* #bytes to read in the case of MDS_READPAGE */ - __u32 generation; + __u32 unused2; /* was "generation" until 2.4.0 */ __u32 suppgid; __u32 eadatasize; __u32 aclsize; @@ -1773,167 +2079,6 @@ 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 */ -#define Q_GETOQUOTA 0x800103 /* get obd quotas */ -#define Q_FINVALIDATE 0x800104 /* invalidate operational quotas */ - -#define Q_TYPEMATCH(id, type) \ - ((id) == (type) || (id) == UGQUOTA) - -#define Q_TYPESET(oqc, type) Q_TYPEMATCH((oqc)->qc_type, type) - -#define Q_GETOCMD(oqc) \ - ((oqc)->qc_cmd == Q_GETOINFO || (oqc)->qc_cmd == Q_GETOQUOTA) - -#define QCTL_COPY(out, in) \ -do { \ - Q_COPY(out, in, qc_cmd); \ - Q_COPY(out, in, qc_type); \ - Q_COPY(out, in, qc_id); \ - Q_COPY(out, in, qc_stat); \ - Q_COPY(out, in, qc_dqinfo); \ - Q_COPY(out, in, qc_dqblk); \ -} while (0) - -struct obd_quotactl { - __u32 qc_cmd; - __u32 qc_type; - __u32 qc_id; - __u32 qc_stat; - struct obd_dqinfo qc_dqinfo; - struct obd_dqblk qc_dqblk; -}; - -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 - * acquire 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; - __u64 qaq_bunit_sz; - __u64 qaq_iunit_sz; - __u64 padding1; -}; -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 */ -#define LQUOTA_FLAGS_ADJBLK 4UL /* adjust the block qunit size */ -#define LQUOTA_FLAGS_ADJINO 8UL /* adjust the inode qunit size */ -#define LQUOTA_FLAGS_CHG_QS 16UL /* indicate whether it has capability of - * OBD_CONNECT_CHANGE_QS */ -#define LQUOTA_FLAGS_RECOVERY 32UL /* recovery is going on a uid/gid */ -#define LQUOTA_FLAGS_SETQUOTA 64UL /* being setquota on a uid/gid */ - -/* flags is specific for quota_adjust_qunit */ -#define LQUOTA_QAQ_CREATE_LQS (1UL << 31) /* when it is set, need create lqs */ - -/* the status of lqs_flags in struct lustre_qunit_size */ -#define LQUOTA_QUNIT_FLAGS (LQUOTA_FLAGS_GRP | LQUOTA_FLAGS_BLK) - -#define QAQ_IS_GRP(qaq) ((qaq)->qaq_flags & LQUOTA_FLAGS_GRP) -#define QAQ_IS_ADJBLK(qaq) ((qaq)->qaq_flags & LQUOTA_FLAGS_ADJBLK) -#define QAQ_IS_ADJINO(qaq) ((qaq)->qaq_flags & LQUOTA_FLAGS_ADJINO) -#define QAQ_IS_CREATE_LQS(qaq) ((qaq)->qaq_flags & LQUOTA_QAQ_CREATE_LQS) - -#define QAQ_SET_GRP(qaq) ((qaq)->qaq_flags |= LQUOTA_FLAGS_GRP) -#define QAQ_SET_ADJBLK(qaq) ((qaq)->qaq_flags |= LQUOTA_FLAGS_ADJBLK) -#define QAQ_SET_ADJINO(qaq) ((qaq)->qaq_flags |= LQUOTA_FLAGS_ADJINO) -#define QAQ_SET_CREATE_LQS(qaq) ((qaq)->qaq_flags |= LQUOTA_QAQ_CREATE_LQS) - /* permissions for md_perm.mp_perm */ enum { CFS_SETUID_PERM = 0x01, @@ -1980,7 +2125,7 @@ struct mdt_rec_setattr { obd_time sa_ctime; __u32 sa_attr_flags; __u32 sa_mode; - __u32 sa_padding_2; + __u32 sa_bias; /* some operation flags */ __u32 sa_padding_3; __u32 sa_padding_4; __u32 sa_padding_5; @@ -2043,17 +2188,15 @@ extern void lustre_swab_mdt_rec_setattr (struct mdt_rec_setattr *sa); * anymore, reserve this flags * just for preventing such bit * to be reused. */ -#define MDS_CREATE_RMT_ACL 01000000000 /* indicate create on remote server - * with default ACL */ -#define MDS_CREATE_SLAVE_OBJ 02000000000 /* indicate create slave object - * actually, this is for create, not - * conflict with other open flags */ + #define MDS_OPEN_LOCK 04000000000 /* This open requires open lock */ #define MDS_OPEN_HAS_EA 010000000000 /* specify object create pattern */ #define MDS_OPEN_HAS_OBJS 020000000000 /* Just set the EA the obj exist */ #define MDS_OPEN_NORESTORE 0100000000000ULL /* Do not restore file at open */ #define MDS_OPEN_NEWSTRIPE 0200000000000ULL /* New stripe needed (restripe or * hsm restore) */ +#define MDS_OPEN_VOLATILE 0400000000000ULL /* File is volatile = created + unlinked */ /* permission for create non-directory file */ #define MAY_CREATE (1 << 7) @@ -2073,16 +2216,18 @@ extern void lustre_swab_mdt_rec_setattr (struct mdt_rec_setattr *sa); #define MAY_RGETFACL (1 << 14) enum { - MDS_CHECK_SPLIT = 1 << 0, - MDS_CROSS_REF = 1 << 1, - MDS_VTX_BYPASS = 1 << 2, - MDS_PERM_BYPASS = 1 << 3, - MDS_SOM = 1 << 4, - MDS_QUOTA_IGNORE = 1 << 5, - 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 */ + MDS_CHECK_SPLIT = 1 << 0, + MDS_CROSS_REF = 1 << 1, + MDS_VTX_BYPASS = 1 << 2, + MDS_PERM_BYPASS = 1 << 3, + MDS_SOM = 1 << 4, + MDS_QUOTA_IGNORE = 1 << 5, + MDS_CLOSE_CLEANUP = 1 << 6, + MDS_KEEP_ORPHAN = 1 << 7, + MDS_RECOV_OPEN = 1 << 8, + MDS_DATA_MODIFIED = 1 << 9, + MDS_CREATE_VOLATILE = 1 << 10, + MDS_OWNEROVERRIDE = 1 << 11, }; /* instance of mdt_reint_rec */ @@ -2111,7 +2256,7 @@ struct mdt_rec_create { * extend cr_flags size without breaking 1.8 compat */ __u32 cr_flags_l; /* for use with open, low 32 bits */ __u32 cr_flags_h; /* for use with open, high 32 bits */ - __u32 cr_padding_3; /* rr_padding_3 */ + __u32 cr_umask; /* umask for create */ __u32 cr_padding_4; /* rr_padding_4 */ }; @@ -2456,14 +2601,11 @@ typedef union { extern void lustre_swab_ldlm_policy_data (ldlm_wire_policy_data_t *d); -/* 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; +union ldlm_gl_desc { + struct ldlm_gl_lquota_desc lquota_desc; }; -extern void lustre_swab_lvb(union ldlm_wire_lvb *); +extern void lustre_swab_gl_desc(union ldlm_gl_desc *); struct ldlm_intent { __u64 opc; @@ -2522,6 +2664,9 @@ struct ldlm_reply { extern void lustre_swab_ldlm_reply (struct ldlm_reply *r); +#define ldlm_flags_to_wire(flags) ((__u32)(flags)) +#define ldlm_flags_from_wire(flags) ((__u64)(flags)) + /* * Opcodes for mountconf (mgs and mgc) */ @@ -2693,6 +2838,14 @@ struct llog_rec_tail { __u32 lrt_index; }; +/* Where data follow just after header */ +#define REC_DATA(ptr) \ + ((void *)((char *)ptr + sizeof(struct llog_rec_hdr))) + +#define REC_DATA_LEN(rec) \ + (rec->lrh_len - sizeof(struct llog_rec_hdr) - \ + sizeof(struct llog_rec_tail)) + struct llog_logid_rec { struct llog_rec_hdr lid_hdr; struct llog_logid lid_id; @@ -2889,7 +3042,8 @@ struct obdo { obd_flag o_flags; obd_count o_nlink; /* brw: checksum */ obd_count o_parent_oid; - obd_count o_misc; /* brw: o_dropped */ + obd_count o_misc; /* brw: o_dropped */ + __u64 o_ioepoch; /* epoch in ost writes */ __u32 o_stripe_idx; /* holds stripe idx */ __u32 o_parent_ver; @@ -2897,16 +3051,16 @@ struct obdo { * locks */ struct llog_cookie o_lcookie; /* destroy: unlink cookie from * MDS */ - __u32 o_uid_h; - __u32 o_gid_h; + __u32 o_uid_h; + __u32 o_gid_h; - __u64 o_data_version; /* getattr: sum of iversion for - * each stripe. - * brw: grant space consumed on - * the client for the write */ - __u64 o_padding_4; - __u64 o_padding_5; - __u64 o_padding_6; + __u64 o_data_version; /* getattr: sum of iversion for + * each stripe. + * brw: grant space consumed on + * the client for the write */ + __u64 o_padding_4; + __u64 o_padding_5; + __u64 o_padding_6; }; #define o_id o_oi.oi_id @@ -3055,6 +3209,7 @@ struct lu_idxpage { * For the time being, we only support fixed-size key & record. */ char lip_entries[0]; }; +extern void lustre_swab_lip_header(struct lu_idxpage *lip); #define LIP_HDR_SIZE (offsetof(struct lu_idxpage, lip_entries)) @@ -3065,67 +3220,6 @@ union lu_page { char lp_array[LU_PAGE_SIZE]; }; -/* this will be used when OBD_CONNECT_CHANGE_QS is set */ -struct qunit_data { - /** - * ID appiles to (uid, gid) - */ - __u32 qd_id; - /** - * LQUOTA_FLAGS_* affect the responding bits - */ - __u32 qd_flags; - /** - * acquire/release count (bytes for block quota) - */ - __u64 qd_count; - /** - * when a master returns the reply to a slave, it will - * contain the current corresponding qunit size - */ - __u64 qd_qunit; - __u64 padding; -}; - -#define QDATA_IS_GRP(qdata) ((qdata)->qd_flags & LQUOTA_FLAGS_GRP) -#define QDATA_IS_BLK(qdata) ((qdata)->qd_flags & LQUOTA_FLAGS_BLK) -#define QDATA_IS_ADJBLK(qdata) ((qdata)->qd_flags & LQUOTA_FLAGS_ADJBLK) -#define QDATA_IS_ADJINO(qdata) ((qdata)->qd_flags & LQUOTA_FLAGS_ADJINO) -#define QDATA_IS_CHANGE_QS(qdata) ((qdata)->qd_flags & LQUOTA_FLAGS_CHG_QS) - -#define QDATA_SET_GRP(qdata) ((qdata)->qd_flags |= LQUOTA_FLAGS_GRP) -#define QDATA_SET_BLK(qdata) ((qdata)->qd_flags |= LQUOTA_FLAGS_BLK) -#define QDATA_SET_ADJBLK(qdata) ((qdata)->qd_flags |= LQUOTA_FLAGS_ADJBLK) -#define QDATA_SET_ADJINO(qdata) ((qdata)->qd_flags |= LQUOTA_FLAGS_ADJINO) -#define QDATA_SET_CHANGE_QS(qdata) ((qdata)->qd_flags |= LQUOTA_FLAGS_CHG_QS) - -#define QDATA_CLR_GRP(qdata) ((qdata)->qd_flags &= ~LQUOTA_FLAGS_GRP) -#define QDATA_CLR_CHANGE_QS(qdata) ((qdata)->qd_flags &= ~LQUOTA_FLAGS_CHG_QS) - -extern void lustre_swab_qdata(struct qunit_data *d); -extern struct qunit_data *quota_get_qdata(void *req, int is_req, int is_exp); -extern int quota_copy_qdata(void *request, struct qunit_data *qdata, - int is_req, int is_exp); - -typedef enum { - QUOTA_DQACQ = 601, - QUOTA_DQREL = 602, - QUOTA_LAST_OPC -} quota_cmd_t; -#define QUOTA_FIRST_OPC QUOTA_DQACQ - -#define QUOTA_REQUEST 1 -#define QUOTA_REPLY 0 -#define QUOTA_EXPORT 1 -#define QUOTA_IMPORT 0 - -/* 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 4 /**< need to acquire extra quota */ - - /* security opcodes */ typedef enum { SEC_CTX_INIT = 801, @@ -3247,6 +3341,145 @@ struct getinfo_fid2path { void lustre_swab_fid2path (struct getinfo_fid2path *gf); +enum { + LAYOUT_INTENT_ACCESS = 0, + LAYOUT_INTENT_READ = 1, + LAYOUT_INTENT_WRITE = 2, + LAYOUT_INTENT_GLIMPSE = 3, + LAYOUT_INTENT_TRUNC = 4, + LAYOUT_INTENT_RELEASE = 5, + LAYOUT_INTENT_RESTORE = 6 +}; + +/* enqueue layout lock with intent */ +struct layout_intent { + __u32 li_opc; /* intent operation for enqueue, read, write etc */ + __u32 li_flags; + __u64 li_start; + __u64 li_end; +}; + +void lustre_swab_layout_intent(struct layout_intent *li); + +/** + * On the wire version of hsm_progress structure. + * + * Contains the userspace hsm_progress and some internal fields. + */ +struct hsm_progress_kernel { + /* Field taken from struct hsm_progress */ + lustre_fid hpk_fid; + __u64 hpk_cookie; + struct hsm_extent hpk_extent; + __u16 hpk_flags; + __u16 hpk_errval; /* positive val */ + __u32 hpk_padding1; + /* Additional fields */ + __u64 hpk_data_version; + __u64 hpk_padding2; +} __attribute__((packed)); + +extern void lustre_swab_hsm_user_state(struct hsm_user_state *hus); +extern void lustre_swab_hsm_current_action(struct hsm_current_action *action); +extern void lustre_swab_hsm_progress_kernel(struct hsm_progress_kernel *hpk); +extern void lustre_swab_hsm_user_state(struct hsm_user_state *hus); +extern void lustre_swab_hsm_user_item(struct hsm_user_item *hui); +extern void lustre_swab_hsm_request(struct hsm_request *hr); + +/** + * These are object update opcode under UPDATE_OBJ, which is currently + * being used by cross-ref operations between MDT. + * + * During the cross-ref operation, the Master MDT, which the client send the + * request to, will disassembly the operation into object updates, then OSP + * will send these updates to the remote MDT to be executed. + * + * Update request format + * magic: UPDATE_BUFFER_MAGIC_V1 + * Count: How many updates in the req. + * bufs[0] : following are packets of object. + * update[0]: + * type: object_update_op, the op code of update + * fid: The object fid of the update. + * lens/bufs: other parameters of the update. + * update[1]: + * type: object_update_op, the op code of update + * fid: The object fid of the update. + * lens/bufs: other parameters of the update. + * .......... + * update[7]: type: object_update_op, the op code of update + * fid: The object fid of the update. + * lens/bufs: other parameters of the update. + * Current 8 maxim updates per object update request. + * + ******************************************************************* + * update reply format: + * + * ur_version: UPDATE_REPLY_V1 + * ur_count: The count of the reply, which is usually equal + * to the number of updates in the request. + * ur_lens: The reply lengths of each object update. + * + * replies: 1st update reply [4bytes_ret: other body] + * 2nd update reply [4bytes_ret: other body] + * ..... + * nth update reply [4bytes_ret: other body] + * + * For each reply of the update, the format would be + * result(4 bytes):Other stuff + */ + +#define UPDATE_MAX_OPS 10 +#define UPDATE_BUFFER_MAGIC_V1 0xBDDE0001 +#define UPDATE_BUFFER_MAGIC UPDATE_BUFFER_MAGIC_V1 +#define UPDATE_BUF_COUNT 8 +enum object_update_op { + OBJ_CREATE = 1, + OBJ_DESTROY = 2, + OBJ_REF_ADD = 3, + OBJ_REF_DEL = 4, + OBJ_ATTR_SET = 5, + OBJ_ATTR_GET = 6, + OBJ_XATTR_SET = 7, + OBJ_XATTR_GET = 8, + OBJ_INDEX_LOOKUP = 9, + OBJ_INDEX_INSERT = 10, + OBJ_INDEX_DELETE = 11, + OBJ_LAST +}; + +struct update { + __u32 u_type; + __u32 u_padding; + struct lu_fid u_fid; + __u32 u_lens[UPDATE_BUF_COUNT]; + __u32 u_bufs[0]; +}; + +struct update_buf { + __u32 ub_magic; + __u32 ub_count; + __u32 ub_bufs[0]; +}; + +#define UPDATE_REPLY_V1 0x00BD0001 +struct update_reply { + __u32 ur_version; + __u32 ur_count; + __u32 ur_lens[0]; +}; + +void lustre_swab_update_buf(struct update_buf *ub); +void lustre_swab_update_reply_buf(struct update_reply *ur); + +/** layout swap request structure + * fid1 and fid2 are in mdt_body + */ +struct mdc_swap_layouts { + __u64 msl_flags; +} __packed; + +void lustre_swab_swap_layouts(struct mdc_swap_layouts *msl); #endif /** @} lustreidl */