X-Git-Url: https://git.whamcloud.com/?p=fs%2Flustre-release.git;a=blobdiff_plain;f=lustre%2Flov%2Flov_pack.c;h=bdcf14674eba2f8ba680a8f4b0ac783ee9dd0053;hp=d2cca20a7b6dd4ca6a7a13bf0f51a24aafe6f705;hb=364ec95f3688ac5cc3195f7f46d0d860844796f9;hpb=3de901fceee79de12a31428bcc6ba3a00f10d1fe diff --git a/lustre/lov/lov_pack.c b/lustre/lov/lov_pack.c index d2cca20..bdcf146 100644 --- a/lustre/lov/lov_pack.c +++ b/lustre/lov/lov_pack.c @@ -1,593 +1,459 @@ -/* -*- mode: c; c-basic-offset: 8; indent-tabs-mode: nil; -*- - * vim:expandtab:shiftwidth=8:tabstop=8: +/* + * GPL HEADER START * - * Copyright (C) 2002, 2003 Cluster File Systems, Inc. - * Author: Andreas Dilger + * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER. * - * This file is part of Lustre, http://www.lustre.org. + * 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. * - * Lustre is free software; you can redistribute it and/or - * modify it under the terms of version 2 of the GNU General Public - * License 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). * - * Lustre 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 for more details. + * You should have received a copy of the GNU General Public License + * version 2 along with this program; If not, see + * http://www.gnu.org/licenses/gpl-2.0.html * - * You should have received a copy of the GNU General Public License - * along with Lustre; if not, write to the Free Software - * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. + * GPL HEADER END + */ +/* + * Copyright (c) 2002, 2010, Oracle and/or its affiliates. All rights reserved. + * Use is subject to license terms. + * + * Copyright (c) 2011, 2015, Intel Corporation. + */ +/* + * This file is part of Lustre, http://www.lustre.org/ + * Lustre is a trademark of Sun Microsystems, Inc. + * + * lustre/lov/lov_pack.c * * (Un)packing of OST/MDS requests * + * Author: Andreas Dilger */ #define DEBUG_SUBSYSTEM S_LOV -#ifndef __KERNEL__ -#include -#endif - -#include -#include -#include -#include -#include + +#include #include +#include +#include +#include +#include +#include + +#include "lov_cl_internal.h" #include "lov_internal.h" -void lov_dump_lmm_v0(int level, struct lov_mds_md_v0 *lmm) +void lov_dump_lmm_common(int level, void *lmmp) { - int i, num_ost, stripe, idx; - - num_ost = le32_to_cpu(lmm->lmm_ost_count); - idx = le32_to_cpu(lmm->lmm_stripe_offset); - CDEBUG(level, "objid "LPX64", magic 0x%08X, ost_count %u\n", - le64_to_cpu(lmm->lmm_object_id), le32_to_cpu(lmm->lmm_magic), - num_ost); - CDEBUG(level,"stripe_size %u, stripe_count %u, stripe_offset %u\n", - le32_to_cpu(lmm->lmm_stripe_size), - le32_to_cpu(lmm->lmm_stripe_count), idx); - for (i = stripe = 0; i < le32_to_cpu(lmm->lmm_ost_count); i++, idx++) { - idx %= num_ost; - if (lmm->lmm_objects[idx].l_object_id == 0) - continue; - CDEBUG(level, "stripe %u idx %u subobj "LPX64"\n", stripe, idx, - le64_to_cpu(lmm->lmm_objects[idx].l_object_id)); - stripe++; - } + struct lov_mds_md *lmm = lmmp; + struct ost_id oi; + + lmm_oi_le_to_cpu(&oi, &lmm->lmm_oi); + CDEBUG(level, "objid "DOSTID", magic 0x%08x, pattern %#x\n", + POSTID(&oi), le32_to_cpu(lmm->lmm_magic), + le32_to_cpu(lmm->lmm_pattern)); + CDEBUG(level, "stripe_size %u, stripe_count %u, layout_gen %u\n", + le32_to_cpu(lmm->lmm_stripe_size), + le16_to_cpu(lmm->lmm_stripe_count), + le16_to_cpu(lmm->lmm_layout_gen)); } -void lov_dump_lmm_v1(int level, struct lov_mds_md_v1 *lmm) +static void lov_dump_lmm_objects(int level, struct lov_ost_data *lod, + int stripe_count) { - struct lov_ost_data_v1 *lod; - int i; - - CDEBUG(level, "objid "LPX64", magic 0x%08X, pattern %#X\n", - le64_to_cpu(lmm->lmm_object_id), le32_to_cpu(lmm->lmm_magic), - le32_to_cpu(lmm->lmm_pattern)); - CDEBUG(level,"stripe_size %u, stripe_count %u\n", - le32_to_cpu(lmm->lmm_stripe_size), - le32_to_cpu(lmm->lmm_stripe_count)); - for (i = 0, lod = lmm->lmm_objects; - i < le32_to_cpu(lmm->lmm_stripe_count); i++, lod++) - CDEBUG(level, "stripe %u idx %u subobj "LPX64"/"LPX64"\n", - i, le32_to_cpu(lod->l_ost_idx), - le64_to_cpu(lod->l_object_gr), - le64_to_cpu(lod->l_object_id)); -} + int i; -#define LMM_ASSERT(test) \ -do { \ - if (!(test)) lov_dump_lmm(D_ERROR, lmm); \ - LASSERT(test); /* so we know what assertion failed */ \ -} while(0) + if (stripe_count > LOV_V1_INSANE_STRIPE_COUNT) { + CDEBUG(level, "bad stripe_count %u > max_stripe_count %u\n", + stripe_count, LOV_V1_INSANE_STRIPE_COUNT); + return; + } -/* Pack LOV object metadata for disk storage. It is packed in LE byte - * order and is opaque to the networking layer. - * - * XXX In the future, this will be enhanced to get the EA size from the - * underlying OSC device(s) to get their EA sizes so we can stack - * LOVs properly. For now lov_mds_md_size() just assumes one obd_id - * per stripe. - */ -int lov_packmd(struct obd_export *exp, struct lov_mds_md **lmmp, - struct lov_stripe_md *lsm) -{ - struct obd_device *obd = class_exp2obd(exp); - struct lov_obd *lov = &obd->u.lov; - struct lov_oinfo *loi; - struct lov_mds_md *lmm; - int stripe_count = lov->desc.ld_tgt_count; - int lmm_size; - int i; - ENTRY; - - if (lsm) { - if (lsm->lsm_magic != LOV_MAGIC) { - CERROR("bad mem LOV MAGIC: 0x%08X != 0x%08X\n", - lsm->lsm_magic, LOV_MAGIC); - RETURN(-EINVAL); - } - stripe_count = lsm->lsm_stripe_count; - } - - /* XXX LOV STACKING call into osc for sizes */ - lmm_size = lov_mds_md_size(stripe_count); - - if (!lmmp) - RETURN(lmm_size); - - if (*lmmp && !lsm) { - stripe_count = le32_to_cpu((*lmmp)->lmm_stripe_count); - OBD_FREE(*lmmp, lov_mds_md_size(stripe_count)); - *lmmp = NULL; - RETURN(0); - } - - if (!*lmmp) { - OBD_ALLOC(*lmmp, lmm_size); - if (!*lmmp) - RETURN(-ENOMEM); - } - - lmm = *lmmp; - lmm->lmm_magic = cpu_to_le32(LOV_MAGIC); /* only write new format */ - - if (!lsm) - RETURN(lmm_size); - - lmm->lmm_object_id = cpu_to_le64(lsm->lsm_object_id); - lmm->lmm_object_gr = cpu_to_le64(lsm->lsm_object_gr); - lmm->lmm_stripe_size = cpu_to_le32(lsm->lsm_stripe_size); - lmm->lmm_stripe_count = cpu_to_le32(stripe_count); - lmm->lmm_pattern = cpu_to_le32(lsm->lsm_pattern); - - for (i = 0, loi = lsm->lsm_oinfo; i < stripe_count; i++, loi++) { - /* XXX LOV STACKING call down to osc_packmd() to do packing */ - LASSERT(loi->loi_id); - lmm->lmm_objects[i].l_object_id = cpu_to_le64(loi->loi_id); - lmm->lmm_objects[i].l_object_gr = cpu_to_le64(loi->loi_gr); - lmm->lmm_objects[i].l_ost_gen = cpu_to_le32(loi->loi_ost_gen); - lmm->lmm_objects[i].l_ost_idx = cpu_to_le32(loi->loi_ost_idx); - } - - RETURN(lmm_size); + for (i = 0; i < stripe_count; ++i, ++lod) { + struct ost_id oi; + + ostid_le_to_cpu(&lod->l_ost_oi, &oi); + CDEBUG(level, "stripe %u idx %u subobj "DOSTID"\n", i, + le32_to_cpu(lod->l_ost_idx), POSTID(&oi)); + } } -int lov_get_stripecnt(struct lov_obd *lov, int stripe_count) +void lov_dump_lmm_v1(int level, struct lov_mds_md_v1 *lmm) { - if (!stripe_count) - stripe_count = lov->desc.ld_default_stripe_count; - if (!stripe_count || stripe_count > lov->desc.ld_active_tgt_count) - stripe_count = lov->desc.ld_active_tgt_count; + lov_dump_lmm_common(level, lmm); + lov_dump_lmm_objects(level, lmm->lmm_objects, + le16_to_cpu(lmm->lmm_stripe_count)); +} - return stripe_count; +void lov_dump_lmm_v3(int level, struct lov_mds_md_v3 *lmm) +{ + lov_dump_lmm_common(level, lmm); + CDEBUG(level,"pool_name "LOV_POOLNAMEF"\n", lmm->lmm_pool_name); + lov_dump_lmm_objects(level, lmm->lmm_objects, + le16_to_cpu(lmm->lmm_stripe_count)); } -static int lov_verify_lmm_v0(struct lov_mds_md_v0 *lmm, int lmm_bytes, - int *stripe_count) +void lov_dump_lmm(int level, void *lmm) { - if (lmm_bytes < sizeof(*lmm)) { - CERROR("lov_mds_md too small: %d, need at least %d\n", - lmm_bytes, (int)sizeof(*lmm)); - return -EINVAL; - } - - *stripe_count = le16_to_cpu(lmm->lmm_stripe_count); - - if (*stripe_count == 0 || - *stripe_count > le32_to_cpu(lmm->lmm_ost_count)) { - CERROR("bad stripe count %d\n", *stripe_count); - lov_dump_lmm_v0(D_WARNING, lmm); - return -EINVAL; - } - - if (lmm_bytes < lov_mds_md_v0_size(*stripe_count)) { - CERROR("LOV EA too small: %d, need %d\n", - lmm_bytes, lov_mds_md_size(*stripe_count)); - lov_dump_lmm_v0(D_WARNING, lmm); - return -EINVAL; - } - - if (lmm->lmm_object_id == 0) { - CERROR("zero object id\n"); - lov_dump_lmm_v0(D_WARNING, lmm); - return -EINVAL; - } - - if (le32_to_cpu(lmm->lmm_stripe_offset) > - le32_to_cpu(lmm->lmm_ost_count)) { - CERROR("stripe offset %d more than number of OSTs %d\n", - le32_to_cpu(lmm->lmm_stripe_offset), - le32_to_cpu(lmm->lmm_ost_count)); - lov_dump_lmm_v0(D_WARNING, lmm); - return -EINVAL; - } - - if (lmm->lmm_stripe_size == 0) { - CERROR("zero stripe size\n"); - lov_dump_lmm_v0(D_WARNING, lmm); - return -EINVAL; - } - - return 0; + int magic; + + magic = le32_to_cpu(((struct lov_mds_md *)lmm)->lmm_magic); + switch (magic) { + case LOV_MAGIC_V1: + lov_dump_lmm_v1(level, (struct lov_mds_md_v1 *)lmm); + break; + case LOV_MAGIC_V3: + lov_dump_lmm_v3(level, (struct lov_mds_md_v3 *)lmm); + break; + default: + CDEBUG(level, "unrecognized lmm_magic %x, assuming %x\n", + magic, LOV_MAGIC_V1); + lov_dump_lmm_common(level, lmm); + break; + } } -static int lov_verify_lmm_v1(struct lov_mds_md_v1 *lmm, int lmm_bytes, - int *stripe_count) +/** + * Pack LOV striping metadata for disk storage format (in little + * endian byte order). + * + * This follows the getxattr() conventions. If \a buf_size is zero + * then return the size needed. If \a buf_size is too small then + * return -ERANGE. Otherwise return the size of the result. + */ +ssize_t lov_lsm_pack_v1v3(const struct lov_stripe_md *lsm, void *buf, + size_t buf_size) { - if (lmm_bytes < sizeof(*lmm)) { - CERROR("lov_mds_md too small: %d, need at least %d\n", - lmm_bytes, (int)sizeof(*lmm)); - return -EINVAL; - } - - if (lmm->lmm_magic != le32_to_cpu(LOV_MAGIC_V1)) { - CERROR("bad disk LOV MAGIC: 0x%08X\n", - le32_to_cpu(*(__u32 *)lmm)); - return -EINVAL; - } - - *stripe_count = le32_to_cpu(lmm->lmm_stripe_count); - - if (*stripe_count == 0) { - CERROR("bad stripe count %d\n", *stripe_count); - lov_dump_lmm_v1(D_WARNING, lmm); - return -EINVAL; - } - - if (lmm_bytes < lov_mds_md_size(*stripe_count)) { - CERROR("LOV EA too small: %d, need %d\n", - lmm_bytes, lov_mds_md_size(*stripe_count)); - lov_dump_lmm_v1(D_WARNING, lmm); - return -EINVAL; - } - - if (lmm->lmm_object_id == 0) { - CERROR("zero object id\n"); - lov_dump_lmm_v1(D_WARNING, lmm); - return -EINVAL; - } - - if (lmm->lmm_pattern != cpu_to_le32(LOV_PATTERN_RAID0)) { - CERROR("bad striping pattern\n"); - lov_dump_lmm_v1(D_WARNING, lmm); - return -EINVAL; - } - - if (lmm->lmm_stripe_size == 0 || - (__u64)le32_to_cpu(lmm->lmm_stripe_size) * *stripe_count > ~0UL) { - CERROR("bad stripe size %u\n", - le32_to_cpu(lmm->lmm_stripe_size)); - lov_dump_lmm_v1(D_WARNING, lmm); - return -EINVAL; - } - - return 0; + struct lov_mds_md_v1 *lmmv1 = buf; + struct lov_mds_md_v3 *lmmv3 = buf; + struct lov_ost_data_v1 *lmm_objects; + size_t lmm_size; + unsigned int i; + ENTRY; + + lmm_size = lov_mds_md_size(lsm->lsm_entries[0]->lsme_stripe_count, + lsm->lsm_magic); + if (buf_size == 0) + RETURN(lmm_size); + + if (buf_size < lmm_size) + RETURN(-ERANGE); + + /* lmmv1 and lmmv3 point to the same struct and have the + * same first fields + */ + lmmv1->lmm_magic = cpu_to_le32(lsm->lsm_magic); + lmm_oi_cpu_to_le(&lmmv1->lmm_oi, &lsm->lsm_oi); + lmmv1->lmm_stripe_size = cpu_to_le32( + lsm->lsm_entries[0]->lsme_stripe_size); + lmmv1->lmm_stripe_count = cpu_to_le16( + lsm->lsm_entries[0]->lsme_stripe_count); + lmmv1->lmm_pattern = cpu_to_le32(lsm->lsm_entries[0]->lsme_pattern); + lmmv1->lmm_layout_gen = cpu_to_le16(lsm->lsm_layout_gen); + + if (lsm->lsm_magic == LOV_MAGIC_V3) { + CLASSERT(sizeof(lsm->lsm_entries[0]->lsme_pool_name) == + sizeof(lmmv3->lmm_pool_name)); + strlcpy(lmmv3->lmm_pool_name, + lsm->lsm_entries[0]->lsme_pool_name, + sizeof(lmmv3->lmm_pool_name)); + lmm_objects = lmmv3->lmm_objects; + } else { + lmm_objects = lmmv1->lmm_objects; + } + + if (lsm->lsm_is_released) + RETURN(lmm_size); + + for (i = 0; i < lsm->lsm_entries[0]->lsme_stripe_count; i++) { + struct lov_oinfo *loi = lsm->lsm_entries[0]->lsme_oinfo[i]; + + ostid_cpu_to_le(&loi->loi_oi, &lmm_objects[i].l_ost_oi); + lmm_objects[i].l_ost_gen = cpu_to_le32(loi->loi_ost_gen); + lmm_objects[i].l_ost_idx = cpu_to_le32(loi->loi_ost_idx); + } + + RETURN(lmm_size); } -static int lov_verify_lmm(void *lmm, int lmm_bytes, int *stripe_count) +ssize_t lov_lsm_pack(const struct lov_stripe_md *lsm, void *buf, + size_t buf_size) { - switch (le32_to_cpu(*(__u32 *)lmm)) { - case LOV_MAGIC_V1: - return lov_verify_lmm_v1(lmm, lmm_bytes, stripe_count); - case LOV_MAGIC_V0: - return lov_verify_lmm_v0(lmm, lmm_bytes, stripe_count); - default: - CERROR("bad disk LOV MAGIC: 0x%08X\n", - le32_to_cpu(*(__u32 *)lmm)); - return -EINVAL; - } + struct lov_comp_md_v1 *lcmv1 = buf; + struct lov_comp_md_entry_v1 *lcme; + struct lov_ost_data_v1 *lmm_objects; + size_t lmm_size; + unsigned int entry; + unsigned int offset; + unsigned int size; + unsigned int i; + ENTRY; + + if (lsm->lsm_magic == LOV_MAGIC_V1 || lsm->lsm_magic == LOV_MAGIC_V3) + return lov_lsm_pack_v1v3(lsm, buf, buf_size); + + lmm_size = lov_comp_md_size(lsm); + if (buf_size == 0) + RETURN(lmm_size); + + if (buf_size < lmm_size) + RETURN(-ERANGE); + + lcmv1->lcm_magic = cpu_to_le32(lsm->lsm_magic); + lcmv1->lcm_size = cpu_to_le32(lmm_size); + lcmv1->lcm_layout_gen = cpu_to_le32(lsm->lsm_layout_gen); + lcmv1->lcm_entry_count = cpu_to_le16(lsm->lsm_entry_count); + + offset = sizeof(*lcmv1) + sizeof(*lcme) * lsm->lsm_entry_count; + + for (entry = 0; entry < lsm->lsm_entry_count; entry++) { + struct lov_stripe_md_entry *lsme; + struct lov_mds_md *lmm; + __u16 stripecnt; + + lsme = lsm->lsm_entries[entry]; + lcme = &lcmv1->lcm_entries[entry]; + + lcme->lcme_id = cpu_to_le32(lsme->lsme_id); + lcme->lcme_flags = cpu_to_le32(lsme->lsme_flags); + lcme->lcme_extent.e_start = + cpu_to_le64(lsme->lsme_extent.e_start); + lcme->lcme_extent.e_end = + cpu_to_le64(lsme->lsme_extent.e_end); + lcme->lcme_offset = cpu_to_le32(offset); + + lmm = (struct lov_mds_md *)((char *)lcmv1 + offset); + lmm->lmm_magic = cpu_to_le32(lsme->lsme_magic); + /* lmm->lmm_oi not set */ + lmm->lmm_pattern = cpu_to_le32(lsme->lsme_pattern); + lmm->lmm_stripe_size = cpu_to_le32(lsme->lsme_stripe_size); + lmm->lmm_stripe_count = cpu_to_le16(lsme->lsme_stripe_count); + lmm->lmm_layout_gen = cpu_to_le16(lsme->lsme_layout_gen); + + if (lsme->lsme_magic == LOV_MAGIC_V3) { + struct lov_mds_md_v3 *lmmv3 = + (struct lov_mds_md_v3 *)lmm; + + strlcpy(lmmv3->lmm_pool_name, lsme->lsme_pool_name, + sizeof(lmmv3->lmm_pool_name)); + lmm_objects = lmmv3->lmm_objects; + } else { + lmm_objects = + ((struct lov_mds_md_v1 *)lmm)->lmm_objects; + } + + if (lsme_inited(lsme) && + !(lsme->lsme_pattern & LOV_PATTERN_F_RELEASED)) + stripecnt = lsme->lsme_stripe_count; + else + stripecnt = 0; + + for (i = 0; i < stripecnt; i++) { + struct lov_oinfo *loi = lsme->lsme_oinfo[i]; + + ostid_cpu_to_le(&loi->loi_oi, &lmm_objects[i].l_ost_oi); + lmm_objects[i].l_ost_gen = + cpu_to_le32(loi->loi_ost_gen); + lmm_objects[i].l_ost_idx = + cpu_to_le32(loi->loi_ost_idx); + } + + size = lov_mds_md_size(stripecnt, lsme->lsme_magic); + lcme->lcme_size = cpu_to_le32(size); + offset += size; + } /* for each layout component */ + + RETURN(lmm_size); } -int lov_alloc_memmd(struct lov_stripe_md **lsmp, int stripe_count, int pattern) +/* Find the max stripecount we should use */ +__u16 lov_get_stripecnt(struct lov_obd *lov, __u32 magic, __u16 stripe_count) { - int lsm_size = lov_stripe_md_size(stripe_count); - struct lov_oinfo *loi; - int i; + __u32 max_stripes = LOV_MAX_STRIPE_COUNT_OLD; - OBD_ALLOC(*lsmp, lsm_size); - if (!*lsmp) - return -ENOMEM; + if (!stripe_count) + stripe_count = lov->desc.ld_default_stripe_count; + if (stripe_count > lov->desc.ld_active_tgt_count) + stripe_count = lov->desc.ld_active_tgt_count; + if (!stripe_count) + stripe_count = 1; - (*lsmp)->lsm_magic = LOV_MAGIC; - (*lsmp)->lsm_stripe_count = stripe_count; - (*lsmp)->lsm_maxbytes = LUSTRE_STRIPE_MAXBYTES * stripe_count; - (*lsmp)->lsm_xfersize = PTLRPC_MTU * stripe_count; - (*lsmp)->lsm_pattern = pattern; - (*lsmp)->lsm_oinfo[0].loi_ost_idx = ~0; + /* stripe count is based on whether ldiskfs can handle + * larger EA sizes */ + if (lov->lov_ocd.ocd_connect_flags & OBD_CONNECT_MAX_EASIZE && + lov->lov_ocd.ocd_max_easize) + max_stripes = lov_mds_md_max_stripe_count( + lov->lov_ocd.ocd_max_easize, magic); - for (i = 0, loi = (*lsmp)->lsm_oinfo; i < stripe_count; i++, loi++) - loi_init(loi); + if (stripe_count > max_stripes) + stripe_count = max_stripes; - return lsm_size; + return stripe_count; } -void lov_free_memmd(struct lov_stripe_md **lsmp) +int lov_free_memmd(struct lov_stripe_md **lsmp) { - OBD_FREE(*lsmp, lov_stripe_md_size((*lsmp)->lsm_stripe_count)); - *lsmp = NULL; -} + struct lov_stripe_md *lsm = *lsmp; + int refc; -int lov_unpackmd_v0(struct lov_obd *lov, struct lov_stripe_md *lsm, - struct lov_mds_md_v0 *lmm) -{ - struct lov_oinfo *loi; - int i, ost_offset, ost_count; - - lsm->lsm_object_id = le64_to_cpu(lmm->lmm_object_id); - /* lsm->lsm_object_gr = 0; implicit */ - lsm->lsm_stripe_size = le32_to_cpu(lmm->lmm_stripe_size); - lsm->lsm_xfersize = lsm->lsm_stripe_size * lsm->lsm_stripe_count; - lsm->lsm_pattern = LOV_PATTERN_RAID0; - ost_offset = le32_to_cpu(lmm->lmm_stripe_offset); - ost_count = le16_to_cpu(lmm->lmm_ost_count); - - for (i = 0, loi = lsm->lsm_oinfo; i < ost_count; i++, ost_offset++) { - ost_offset %= ost_count; - - if (!lmm->lmm_objects[ost_offset].l_object_id) - continue; - - loi->loi_id = - le64_to_cpu(lmm->lmm_objects[ost_offset].l_object_id); - /* loi->loi_gr = 0; implicit */ - loi->loi_ost_idx = ost_offset; - /* loi->loi_ost_gen = 0; implicit */ - loi++; - } - - if (loi - lsm->lsm_oinfo != lsm->lsm_stripe_count) { - CERROR("missing objects in lmm struct\n"); - lov_dump_lmm_v0(D_WARNING, lmm); - return -EINVAL; - } - - return 0; -} + *lsmp = NULL; + refc = atomic_dec_return(&lsm->lsm_refc); + LASSERT(refc >= 0); + if (refc == 0) + lsm_free(lsm); -int lov_unpackmd_v1(struct lov_obd *lov, struct lov_stripe_md *lsm, - struct lov_mds_md_v1 *lmm) -{ - struct lov_oinfo *loi; - int i; - - lsm->lsm_object_id = le64_to_cpu(lmm->lmm_object_id); - lsm->lsm_object_gr = le64_to_cpu(lmm->lmm_object_gr); - lsm->lsm_stripe_size = le32_to_cpu(lmm->lmm_stripe_size); - lsm->lsm_pattern = le32_to_cpu(lmm->lmm_pattern); - lsm->lsm_xfersize = lsm->lsm_stripe_size * lsm->lsm_stripe_count; - - for (i = 0, loi = lsm->lsm_oinfo; i < lsm->lsm_stripe_count; i++) { - /* XXX LOV STACKING call down to osc_unpackmd() */ - loi->loi_id = le64_to_cpu(lmm->lmm_objects[i].l_object_id); - loi->loi_gr = le64_to_cpu(lmm->lmm_objects[i].l_object_gr); - loi->loi_ost_idx = le32_to_cpu(lmm->lmm_objects[i].l_ost_idx); - loi->loi_ost_gen = le32_to_cpu(lmm->lmm_objects[i].l_ost_gen); - if (loi->loi_ost_idx > lov->desc.ld_tgt_count) { - CERROR("OST index %d more than OST count %d\n", - loi->loi_ost_idx, lov->desc.ld_tgt_count); - lov_dump_lmm_v1(D_WARNING, lmm); - return -EINVAL; - } - loi++; - } - - return 0; + return refc; } /* Unpack LOV object metadata from disk storage. It is packed in LE byte * order and is opaque to the networking layer. */ -int lov_unpackmd(struct obd_export *exp, struct lov_stripe_md **lsmp, - struct lov_mds_md *lmm, int lmm_bytes) +struct lov_stripe_md *lov_unpackmd(struct lov_obd *lov, void *buf, + size_t buf_size) { - struct obd_device *obd = class_exp2obd(exp); - struct lov_obd *lov = &obd->u.lov; - int rc = 0, stripe_count, lsm_size; - ENTRY; - - /* If passed an MDS struct use values from there, otherwise defaults */ - if (lmm) { - rc = lov_verify_lmm(lmm, lmm_bytes, &stripe_count); - if (rc) - RETURN(rc); - } else { - stripe_count = lov_get_stripecnt(lov, 0); - } - - /* If we aren't passed an lsmp struct, we just want the size */ - if (!lsmp) - /* XXX LOV STACKING call into osc for sizes */ - RETURN(lov_stripe_md_size(stripe_count)); - - /* If we are passed an allocated struct but nothing to unpack, free */ - if (*lsmp && !lmm) { - lov_free_memmd(lsmp); - RETURN(0); - } - - lsm_size = lov_alloc_memmd(lsmp, stripe_count, LOV_PATTERN_RAID0); - if (lsm_size < 0) - RETURN(lsm_size); - - /* If we are passed a pointer but nothing to unpack, we only alloc */ - if (!lmm) - RETURN(lsm_size); - - switch (le32_to_cpu(lmm->lmm_magic)) { - case LOV_MAGIC_V1: - rc = lov_unpackmd_v1(lov, *lsmp, lmm); - break; - case LOV_MAGIC_V0: - rc = lov_unpackmd_v0(lov, *lsmp, (void *)lmm); - break; - } - - if (rc) { - lov_free_memmd(lsmp); - RETURN(rc); - } - - RETURN(lsm_size); -} + const struct lsm_operations *op; + struct lov_stripe_md *lsm; + u32 magic; + ENTRY; -/* Configure object striping information on a new file. - * - * @lmmu is a pointer to a user struct with one or more of the fields set to - * indicate the application preference: lmm_stripe_count, lmm_stripe_size, - * lmm_stripe_offset, and lmm_stripe_pattern. lmm_magic must be LOV_MAGIC. - * @lsmp is a pointer to an in-core stripe MD that needs to be filled in. - */ -int lov_setstripe(struct obd_export *exp, struct lov_stripe_md **lsmp, - struct lov_user_md *lump) -{ - struct obd_device *obd = class_exp2obd(exp); - struct lov_obd *lov = &obd->u.lov; - struct lov_user_md lum; - int stripe_count; - int rc; - ENTRY; - - rc = copy_from_user(&lum, lump, sizeof(lum)); - if (rc) - RETURN(-EFAULT); - - if (lum.lmm_magic != LOV_USER_MAGIC) { - CDEBUG(D_IOCTL, "bad userland LOV MAGIC: %#08x != %#08x\n", - lum.lmm_magic, LOV_USER_MAGIC); - RETURN(-EINVAL); - } - - if (lum.lmm_pattern == 0) { - lum.lmm_pattern = lov->desc.ld_pattern ? - lov->desc.ld_pattern : LOV_PATTERN_RAID0; - } - - if (lum.lmm_pattern != LOV_PATTERN_RAID0) { - CDEBUG(D_IOCTL, "bad userland stripe pattern: %#x\n", - lum.lmm_pattern); - RETURN(-EINVAL); - } - - if (lum.lmm_stripe_size & (PAGE_SIZE - 1)) { - CDEBUG(D_IOCTL, "stripe size %u not multiple of %lu\n", - lum.lmm_stripe_size, PAGE_SIZE); - RETURN(-EINVAL); - } - - if ((lum.lmm_stripe_offset >= lov->desc.ld_active_tgt_count) && - (lum.lmm_stripe_offset != (typeof(lum.lmm_stripe_offset))(-1))) { - CDEBUG(D_IOCTL, "stripe offset %u > number of active OSTs %u\n", - lum.lmm_stripe_offset, lov->desc.ld_active_tgt_count); - RETURN(-EINVAL); - } - stripe_count = lov_get_stripecnt(lov, lum.lmm_stripe_count); - - if ((__u64)lum.lmm_stripe_size * stripe_count > ~0UL) { - CDEBUG(D_IOCTL, "stripe width %ux%u > %lu on 32-bit system\n", - lum.lmm_stripe_size, (int)lum.lmm_stripe_count, ~0UL); - RETURN(-EINVAL); - } - - rc = lov_alloc_memmd(lsmp, stripe_count, lum.lmm_pattern); - - if (rc < 0) - RETURN(rc); - - (*lsmp)->lsm_oinfo[0].loi_ost_idx = lum.lmm_stripe_offset; - (*lsmp)->lsm_stripe_size = lum.lmm_stripe_size; - (*lsmp)->lsm_xfersize = lum.lmm_stripe_size * stripe_count; - - RETURN(0); -} + if (buf_size < sizeof(magic)) + RETURN(ERR_PTR(-EINVAL)); -int lov_setea(struct obd_export *exp, struct lov_stripe_md **lsmp, - struct lov_user_md *lump) -{ - int i; - int rc; - struct obd_export *oexp; - struct lov_obd *lov = &exp->exp_obd->u.lov; - obd_id last_id = 0; - - for (i = 0; i < lump->lmm_stripe_count; i++) { - __u32 len = sizeof(last_id); - oexp = lov->tgts[lump->lmm_objects[i].l_ost_idx].ltd_exp; - rc = obd_get_info(oexp, strlen("last_id"), "last_id", - &len, &last_id); - if (rc) - RETURN(rc); - if (lump->lmm_objects[i].l_object_id > last_id) { - CERROR("Setting EA for object > than last id on " - "ost idx %d "LPD64" > "LPD64" \n", - lump->lmm_objects[i].l_ost_idx, - lump->lmm_objects[i].l_object_id, last_id); - RETURN(-EINVAL); - } - } - - rc = lov_setstripe(exp, lsmp, lump); - if (rc) - RETURN(rc); - - for (i = 0; i < lump->lmm_stripe_count; i++) { - (*lsmp)->lsm_oinfo[i].loi_ost_idx = - lump->lmm_objects[i].l_ost_idx; - (*lsmp)->lsm_oinfo[i].loi_id = lump->lmm_objects[i].l_object_id; - (*lsmp)->lsm_oinfo[i].loi_gr = lump->lmm_objects[i].l_object_gr; - } - RETURN(0); -} + magic = le32_to_cpu(*(u32 *)buf); + op = lsm_op_find(magic); + if (op == NULL) + RETURN(ERR_PTR(-EINVAL)); + lsm = op->lsm_unpackmd(lov, buf, buf_size); + + RETURN(lsm); +} /* Retrieve object striping information. * * @lump is a pointer to an in-core struct with lmm_ost_count indicating * the maximum number of OST indices which will fit in the user buffer. * lmm_magic must be LOV_USER_MAGIC. + * + * If @size > 0, User specified limited buffer size, usually the buffer is from + * ll_lov_setstripe(), and the buffer can only hold basic layout template info. */ -int lov_getstripe(struct obd_export *exp, struct lov_stripe_md *lsm, - struct lov_user_md *lump) +int lov_getstripe(const struct lu_env *env, struct lov_object *obj, + struct lov_stripe_md *lsm, struct lov_user_md __user *lump, + size_t size) { - struct lov_user_md lum; - struct lov_mds_md *lmmk = NULL; - int rc, lmm_size; - ENTRY; - - if (!lsm) - RETURN(-ENODATA); - - rc = copy_from_user(&lum, lump, sizeof(lum)); - if (rc) - RETURN(-EFAULT); - - if (lum.lmm_magic != LOV_USER_MAGIC) - RETURN(-EINVAL); - - rc = lov_packmd(exp, &lmmk, lsm); - if (rc < 0) - RETURN(rc); - lmm_size = rc; - rc = 0; - - /* FIXME: Bug 1185 - copy fields properly when structs change */ - LASSERT(sizeof(lum) == sizeof(*lmmk)); - LASSERT(sizeof(lum.lmm_objects[0]) == sizeof(lmmk->lmm_objects[0])); - - /* User wasn't expecting this many OST entries */ - if (lum.lmm_stripe_count == 0) { - if (copy_to_user(lump, lmmk, sizeof(lum))) - rc = -EFAULT; - } else if (lum.lmm_stripe_count < lmmk->lmm_stripe_count) { - rc = -EOVERFLOW; - } else if (copy_to_user(lump, lmmk, lmm_size)) { - rc = -EFAULT; - } - - obd_free_diskmd(exp, &lmmk); - - RETURN(rc); + /* we use lov_user_md_v3 because it is larger than lov_user_md_v1 */ + struct lov_mds_md *lmmk, *lmm; + struct lov_user_md_v1 lum; + size_t lmmk_size; + ssize_t lmm_size, lum_size = 0; + static bool printed; + int rc = 0; + ENTRY; + + if (lsm->lsm_magic != LOV_MAGIC_V1 && lsm->lsm_magic != LOV_MAGIC_V3 && + lsm->lsm_magic != LOV_MAGIC_COMP_V1) { + CERROR("bad LSM MAGIC: 0x%08X != 0x%08X nor 0x%08X\n", + lsm->lsm_magic, LOV_MAGIC_V1, LOV_MAGIC_V3); + GOTO(out, rc = -EIO); + } + + if (!printed) { + LCONSOLE_WARN("%s: using old ioctl(LL_IOC_LOV_GETSTRIPE) on " + DFID", use llapi_layout_get_by_path()\n", + current->comm, + PFID(&obj->lo_cl.co_lu.lo_header->loh_fid)); + printed = true; + } + + lmmk_size = lov_comp_md_size(lsm); + + OBD_ALLOC_LARGE(lmmk, lmmk_size); + if (lmmk == NULL) + GOTO(out, rc = -ENOMEM); + + lmm_size = lov_lsm_pack(lsm, lmmk, lmmk_size); + if (lmm_size < 0) + GOTO(out_free, rc = lmm_size); + + if (cpu_to_le32(LOV_MAGIC) != LOV_MAGIC) { + if (lmmk->lmm_magic == cpu_to_le32(LOV_MAGIC_V1) || + lmmk->lmm_magic == cpu_to_le32(LOV_MAGIC_V3)) { + lustre_swab_lov_mds_md(lmmk); + lustre_swab_lov_user_md_objects( + (struct lov_user_ost_data *)lmmk->lmm_objects, + lmmk->lmm_stripe_count); + } else if (lmmk->lmm_magic == cpu_to_le32(LOV_MAGIC_COMP_V1)) { + lustre_swab_lov_comp_md_v1( + (struct lov_comp_md_v1 *)lmmk); + } + } + + /* Legacy appication passes limited buffer, we need to figure out + * the user buffer size by the passed in lmm_stripe_count. */ + if (copy_from_user(&lum, lump, sizeof(struct lov_user_md_v1))) + GOTO(out_free, rc = -EFAULT); + + if (lum.lmm_magic == LOV_USER_MAGIC_V1 || + lum.lmm_magic == LOV_USER_MAGIC_V3) + lum_size = lov_user_md_size(lum.lmm_stripe_count, + lum.lmm_magic); + + if (lum_size != 0) { + struct lov_mds_md *comp_md = lmmk; + + /* Legacy app (ADIO for instance) treats the layout as V1/V3 + * blindly, we'd return a reasonable V1/V3 for them. */ + if (lmmk->lmm_magic == LOV_MAGIC_COMP_V1) { + struct lov_comp_md_v1 *comp_v1; + struct cl_object *cl_obj; + struct cl_attr attr; + int i; + + attr.cat_size = 0; + cl_obj = cl_object_top(&obj->lo_cl); + cl_object_attr_get(env, cl_obj, &attr); + + /* return the last instantiated component if file size + * is non-zero, otherwise, return the last component.*/ + comp_v1 = (struct lov_comp_md_v1 *)lmmk; + i = attr.cat_size == 0 ? comp_v1->lcm_entry_count : 0; + for (; i < comp_v1->lcm_entry_count; i++) { + if (!(comp_v1->lcm_entries[i].lcme_flags & + LCME_FL_INIT)) + break; + } + if (i > 0) + i--; + comp_md = (struct lov_mds_md *)((char *)comp_v1 + + comp_v1->lcm_entries[i].lcme_offset); + } + + lmm = comp_md; + lmm_size = lum_size; + } else { + lmm = lmmk; + lmm_size = lmmk_size; + } + /** + * User specified limited buffer size, usually the buffer is + * from ll_lov_setstripe(), and the buffer can only hold basic + * layout template info. + */ + if (size == 0 || size > lmm_size) + size = lmm_size; + if (copy_to_user(lump, lmm, size)) + GOTO(out_free, rc = -EFAULT); + +out_free: + OBD_FREE_LARGE(lmmk, lmmk_size); +out: + RETURN(rc); }