-/* -*- mode: c; c-basic-offset: 8; indent-tabs-mode: nil; -*-
- * vim:expandtab:shiftwidth=8:tabstop=8:
- *
+/*
* GPL HEADER START
*
* DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
*
* You should have received a copy of the GNU General Public License
* version 2 along with this program; If not, see
- * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
- *
- * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
- * CA 95054 USA or visit www.sun.com if you need additional information or
- * have any questions.
+ * http://www.gnu.org/licenses/gpl-2.0.html
*
* GPL HEADER END
*/
* Copyright (c) 2002, 2010, Oracle and/or its affiliates. All rights reserved.
* Use is subject to license terms.
*
- * Copyright (c) 2011, 2012, Whamcloud, Inc.
+ * Copyright (c) 2011, 2017, Intel Corporation.
*/
/*
* This file is part of Lustre, http://www.lustre.org/
*/
#define DEBUG_SUBSYSTEM S_LOV
-#ifndef __KERNEL__
-#include <liblustre.h>
-#endif
#include <lustre_net.h>
+#include <lustre_swab.h>
#include <obd.h>
-#include <obd_lov.h>
#include <obd_class.h>
#include <obd_support.h>
-#include <lustre/lustre_user.h>
+#include "lov_cl_internal.h"
#include "lov_internal.h"
-static void lov_dump_lmm_common(int level, void *lmmp)
+void lov_dump_lmm_common(int level, void *lmmp)
{
- struct lov_mds_md *lmm = lmmp;
-
- CDEBUG(level, "objid "LPX64", magic 0x%08x, pattern %#x\n",
- (__u64)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, 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));
+ struct lov_mds_md *lmm = lmmp;
+ struct ost_id oi;
+
+ lmm_oi_le_to_cpu(&oi, &lmm->lmm_oi);
+ CDEBUG_LIMIT(level, "objid "DOSTID", magic 0x%08x, pattern %#x\n",
+ POSTID(&oi), le32_to_cpu(lmm->lmm_magic),
+ le32_to_cpu(lmm->lmm_pattern));
+ CDEBUG_LIMIT(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));
}
static void lov_dump_lmm_objects(int level, struct lov_ost_data *lod,
- int stripe_count)
+ int stripe_count)
{
- int i;
-
- 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);
- }
-
- for (i = 0; i < stripe_count; ++i, ++lod) {
- CDEBUG(level, "stripe %u idx %u subobj "LPX64"/"LPX64"\n", i,
- le32_to_cpu(lod->l_ost_idx),
- (__u64)le64_to_cpu(lod->l_object_seq),
- (__u64)le64_to_cpu(lod->l_object_id));
- }
+ int i;
+
+ if (stripe_count > LOV_V1_INSANE_STRIPE_COUNT) {
+ CDEBUG_LIMIT(level,
+ "bad stripe_count %u > max_stripe_count %u\n",
+ stripe_count, LOV_V1_INSANE_STRIPE_COUNT);
+ return;
+ }
+
+ for (i = 0; i < stripe_count; ++i, ++lod) {
+ struct ost_id oi;
+
+ ostid_le_to_cpu(&lod->l_ost_oi, &oi);
+ CDEBUG_LIMIT(level, "stripe %u idx %u subobj "DOSTID"\n", i,
+ le32_to_cpu(lod->l_ost_idx), POSTID(&oi));
+ }
}
void lov_dump_lmm_v1(int level, struct lov_mds_md_v1 *lmm)
{
- lov_dump_lmm_common(level, lmm);
- lov_dump_lmm_objects(level, lmm->lmm_objects,
- le16_to_cpu(lmm->lmm_stripe_count));
+ lov_dump_lmm_common(level, lmm);
+ lov_dump_lmm_objects(level, lmm->lmm_objects,
+ le16_to_cpu(lmm->lmm_stripe_count));
}
-void lov_dump_lmm_v3(int level, struct lov_mds_md_v3 *lmm)
+/**
+ * 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.
+ */
+static ssize_t lov_lsm_pack_v1v3(const struct lov_stripe_md *lsm, void *buf,
+ size_t buf_size)
{
- 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));
+ 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) {
+ BUILD_BUG_ON(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);
}
-void lov_dump_lmm(int level, void *lmm)
+static ssize_t lov_lsm_pack_foreign(const struct lov_stripe_md *lsm, void *buf,
+ size_t buf_size)
{
- int magic;
-
- magic = ((struct lov_mds_md_v1 *)(lmm))->lmm_magic;
- switch (magic) {
- case LOV_MAGIC_V1:
- return lov_dump_lmm_v1(level, (struct lov_mds_md_v1 *)(lmm));
- case LOV_MAGIC_V3:
- return lov_dump_lmm_v3(level, (struct lov_mds_md_v3 *)(lmm));
- default:
- CERROR("Cannot recognize lmm_magic %x", magic);
- }
- return;
-}
+ struct lov_foreign_md *lfm = buf;
+ size_t lfm_size;
-#define LMM_ASSERT(test) \
-do { \
- if (!(test)) lov_dump_lmm(D_ERROR, lmm); \
- LASSERT(test); /* so we know what assertion failed */ \
-} while(0)
+ lfm_size = lsm->lsm_foreign_size;
-/* 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_mds_md_v1 *lmmv1;
- struct lov_mds_md_v3 *lmmv3;
- __u16 stripe_count;
- struct lov_ost_data_v1 *lmm_objects;
- int lmm_size, lmm_magic;
- int i;
- ENTRY;
-
- if (lsm) {
- lmm_magic = lsm->lsm_magic;
- } else {
- if (lmmp && *lmmp)
- lmm_magic = le32_to_cpu((*lmmp)->lmm_magic);
- else
- /* lsm == NULL and lmmp == NULL */
- lmm_magic = LOV_MAGIC;
- }
-
- if ((lmm_magic != LOV_MAGIC_V1) &&
- (lmm_magic != LOV_MAGIC_V3)) {
- CERROR("bad mem LOV MAGIC: 0x%08X != 0x%08X nor 0x%08X\n",
- lmm_magic, LOV_MAGIC_V1, LOV_MAGIC_V3);
- RETURN(-EINVAL);
-
- }
-
- if (lsm) {
- /* If we are just sizing the EA, limit the stripe count
- * to the actual number of OSTs in this filesystem. */
- if (!lmmp) {
- stripe_count = lov_get_stripecnt(lov, lmm_magic,
- lsm->lsm_stripe_count);
- lsm->lsm_stripe_count = stripe_count;
- } else {
- stripe_count = lsm->lsm_stripe_count;
- }
- } else {
- /* No need to allocate more than maximum supported stripes.
- * Anyway, this is pretty inaccurate since ld_tgt_count now
- * represents max index and we should rely on the actual number
- * of OSTs instead */
- stripe_count = lov_mds_md_stripecnt(lov->lov_ocd.ocd_max_easize,
- lmm_magic);
- if (stripe_count > lov->desc.ld_tgt_count)
- stripe_count = lov->desc.ld_tgt_count;
- }
-
- /* XXX LOV STACKING call into osc for sizes */
- lmm_size = lov_mds_md_size(stripe_count, lmm_magic);
-
- if (!lmmp)
- RETURN(lmm_size);
-
- if (*lmmp && !lsm) {
- stripe_count = le16_to_cpu((*lmmp)->lmm_stripe_count);
- lmm_size = lov_mds_md_size(stripe_count, lmm_magic);
- OBD_FREE_LARGE(*lmmp, lmm_size);
- *lmmp = NULL;
- RETURN(0);
- }
-
- if (!*lmmp) {
- OBD_ALLOC_LARGE(*lmmp, lmm_size);
- if (!*lmmp)
- RETURN(-ENOMEM);
- }
-
- CDEBUG(D_INFO, "lov_packmd: LOV_MAGIC 0x%08X, lmm_size = %d \n",
- lmm_magic, lmm_size);
-
- lmmv1 = *lmmp;
- lmmv3 = (struct lov_mds_md_v3 *)*lmmp;
- if (lmm_magic == LOV_MAGIC_V3)
- lmmv3->lmm_magic = cpu_to_le32(LOV_MAGIC_V3);
- else
- lmmv1->lmm_magic = cpu_to_le32(LOV_MAGIC_V1);
-
- if (!lsm)
- RETURN(lmm_size);
-
- /* lmmv1 and lmmv3 point to the same struct and have the
- * same first fields
- */
- lmmv1->lmm_object_id = cpu_to_le64(lsm->lsm_object_id);
- lmmv1->lmm_object_seq = cpu_to_le64(lsm->lsm_object_seq);
- lmmv1->lmm_stripe_size = cpu_to_le32(lsm->lsm_stripe_size);
- lmmv1->lmm_stripe_count = cpu_to_le16(stripe_count);
- lmmv1->lmm_pattern = cpu_to_le32(lsm->lsm_pattern);
- lmmv1->lmm_layout_gen = cpu_to_le16(lsm->lsm_layout_gen);
- if (lsm->lsm_magic == LOV_MAGIC_V3) {
- strncpy(lmmv3->lmm_pool_name, lsm->lsm_pool_name,
- LOV_MAXPOOLNAME);
- lmm_objects = lmmv3->lmm_objects;
- } else {
- lmm_objects = lmmv1->lmm_objects;
- }
-
- for (i = 0; i < stripe_count; i++) {
- struct lov_oinfo *loi = lsm->lsm_oinfo[i];
- /* XXX LOV STACKING call down to osc_packmd() to do packing */
- LASSERTF(loi->loi_id, "lmm_oid "LPU64" stripe %u/%u idx %u\n",
- lmmv1->lmm_object_id, i, stripe_count, loi->loi_ost_idx);
- lmm_objects[i].l_object_id = cpu_to_le64(loi->loi_id);
- lmm_objects[i].l_object_seq = cpu_to_le64(loi->loi_seq);
- 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);
-}
+ if (buf_size == 0)
+ RETURN(lfm_size);
-/* Find the max stripecount we should use */
-__u16 lov_get_stripecnt(struct lov_obd *lov, __u32 magic, __u16 stripe_count)
-{
- __u32 max_stripes = LOV_MAX_STRIPE_COUNT_OLD;
-
- 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;
-
- /* 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_stripecnt(lov->lov_ocd.ocd_max_easize,
- magic);
-
- if (stripe_count > max_stripes)
- stripe_count = max_stripes;
-
- return stripe_count;
-}
+ if (buf_size < lfm_size)
+ RETURN(-ERANGE);
+
+ /* full foreign LOV is already avail in its cache
+ * no need to translate format fields to little-endian
+ */
+ memcpy(lfm, lsm_foreign(lsm), lsm->lsm_foreign_size);
+ RETURN(lfm_size);
+}
-static int lov_verify_lmm(void *lmm, int lmm_bytes, __u16 *stripe_count)
+ssize_t lov_lsm_pack(const struct lov_stripe_md *lsm, void *buf,
+ size_t buf_size)
{
- int rc;
-
- if (lsm_op_find(le32_to_cpu(*(__u32 *)lmm)) == NULL) {
- char *buffer;
- int sz;
-
- CERROR("bad disk LOV MAGIC: 0x%08X; dumping LMM (size=%d):\n",
- le32_to_cpu(*(__u32 *)lmm), lmm_bytes);
- sz = lmm_bytes * 2 + 1;
- OBD_ALLOC_LARGE(buffer, sz);
- if (buffer != NULL) {
- int i;
-
- for (i = 0; i < lmm_bytes; i++)
- sprintf(buffer+2*i, "%.2X", ((char *)lmm)[i]);
- buffer[sz] = '\0';
- CERROR("%s\n", buffer);
- OBD_FREE_LARGE(buffer, sz);
- }
- return -EINVAL;
- }
- rc = lsm_op_find(le32_to_cpu(*(__u32 *)lmm))->lsm_lmm_verify(lmm,
- lmm_bytes, stripe_count);
- return rc;
+ 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);
+
+ if (lsm->lsm_magic == LOV_MAGIC_FOREIGN)
+ return lov_lsm_pack_foreign(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_flags = cpu_to_le16(lsm->lsm_flags);
+ lcmv1->lcm_mirror_count = cpu_to_le16(lsm->lsm_mirror_count);
+ 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 stripe_count;
+
+ 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);
+ if (lsme->lsme_flags & LCME_FL_NOSYNC)
+ lcme->lcme_timestamp =
+ cpu_to_le64(lsme->lsme_timestamp);
+ 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))
+ stripe_count = lsme->lsme_stripe_count;
+ else
+ stripe_count = 0;
+
+ for (i = 0; i < stripe_count; 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(stripe_count, 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, __u16 stripe_count,
- int pattern, int magic)
+/* Find the max stripecount we should use */
+__u16 lov_get_stripe_count(struct lov_obd *lov, __u32 magic, __u16 stripe_count)
{
- int i, lsm_size;
- ENTRY;
-
- CDEBUG(D_INFO, "alloc lsm, stripe_count %d\n", stripe_count);
-
- *lsmp = lsm_alloc_plain(stripe_count, &lsm_size);
- if (!*lsmp) {
- CERROR("can't allocate lsmp stripe_count %d\n", stripe_count);
- RETURN(-ENOMEM);
- }
-
- cfs_spin_lock_init(&(*lsmp)->lsm_lock);
- (*lsmp)->lsm_magic = magic;
- (*lsmp)->lsm_stripe_count = stripe_count;
- (*lsmp)->lsm_maxbytes = LUSTRE_STRIPE_MAXBYTES * stripe_count;
- (*lsmp)->lsm_pattern = pattern;
- (*lsmp)->lsm_pool_name[0] = '\0';
- (*lsmp)->lsm_layout_gen = 0;
- (*lsmp)->lsm_oinfo[0]->loi_ost_idx = ~0;
-
- for (i = 0; i < stripe_count; i++)
- loi_init((*lsmp)->lsm_oinfo[i]);
-
- RETURN(lsm_size);
+ __u32 max_stripes = LOV_MAX_STRIPE_COUNT_OLD;
+
+ 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;
+
+ /*
+ * 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);
+
+ if (stripe_count > max_stripes)
+ stripe_count = max_stripes;
+
+ return stripe_count;
}
-void lov_free_memmd(struct lov_stripe_md **lsmp)
+int lov_free_memmd(struct lov_stripe_md **lsmp)
{
- struct lov_stripe_md *lsm = *lsmp;
+ struct lov_stripe_md *lsm = *lsmp;
+ int refc;
- LASSERT(lsm_op_find(lsm->lsm_magic) != NULL);
- lsm_op_find(lsm->lsm_magic)->lsm_free(lsm);
+ *lsmp = NULL;
+ refc = atomic_dec_return(&lsm->lsm_refc);
+ LASSERT(refc >= 0);
+ if (refc == 0)
+ lsm_free(lsm);
- *lsmp = NULL;
+ return refc;
}
-
-/* Unpack LOV object metadata from disk storage. It is packed in LE byte
+/*
+ * 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, lsm_size;
- __u16 stripe_count;
- __u32 magic;
- 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);
- magic = le32_to_cpu(lmm->lmm_magic);
- } else {
- magic = LOV_MAGIC;
- stripe_count = lov_get_stripecnt(lov, magic, 0);
- }
-
- /* If we aren't passed an lsmp struct, we just want the size */
- if (!lsmp) {
- /* XXX LOV STACKING call into osc for sizes */
- LBUG();
- 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,
- magic);
- 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);
-
- LASSERT(lsm_op_find(magic) != NULL);
- rc = lsm_op_find(magic)->lsm_unpackmd(lov, *lsmp, lmm);
- if (rc) {
- lov_free_memmd(lsmp);
- RETURN(rc);
- }
-
- RETURN(lsm_size);
-}
+ const struct lsm_operations *op;
+ struct lov_stripe_md *lsm;
+ u32 magic;
-static int __lov_setstripe(struct obd_export *exp, int max_lmm_size,
- struct lov_stripe_md **lsmp,
- struct lov_user_md *lump)
-{
- struct obd_device *obd = class_exp2obd(exp);
- struct lov_obd *lov = &obd->u.lov;
- char buffer[sizeof(struct lov_user_md_v3)];
- struct lov_user_md_v3 *lumv3 = (struct lov_user_md_v3 *)&buffer[0];
- struct lov_user_md_v1 *lumv1 = (struct lov_user_md_v1 *)&buffer[0];
- int lmm_magic;
- __u16 stripe_count;
- int rc;
- ENTRY;
-
- rc = lov_lum_swab_if_needed(lumv3, &lmm_magic, lump);
- if (rc)
- RETURN(rc);
-
- /* in the rest of the tests, as *lumv1 and lumv3 have the same
- * fields, we use lumv1 to avoid code duplication */
-
- if (lumv1->lmm_pattern == 0) {
- lumv1->lmm_pattern = lov->desc.ld_pattern ?
- lov->desc.ld_pattern : LOV_PATTERN_RAID0;
- }
-
- if (lumv1->lmm_pattern != LOV_PATTERN_RAID0) {
- CDEBUG(D_IOCTL, "bad userland stripe pattern: %#x\n",
- lumv1->lmm_pattern);
- RETURN(-EINVAL);
- }
-
- /* 64kB is the largest common page size we see (ia64), and matches the
- * check in lfs */
- if (lumv1->lmm_stripe_size & (LOV_MIN_STRIPE_SIZE - 1)) {
- CDEBUG(D_IOCTL, "stripe size %u not multiple of %u, fixing\n",
- lumv1->lmm_stripe_size, LOV_MIN_STRIPE_SIZE);
- lumv1->lmm_stripe_size = LOV_MIN_STRIPE_SIZE;
- }
-
- if ((lumv1->lmm_stripe_offset >= lov->desc.ld_tgt_count) &&
- (lumv1->lmm_stripe_offset !=
- (typeof(lumv1->lmm_stripe_offset))(-1))) {
- CDEBUG(D_IOCTL, "stripe offset %u > number of OSTs %u\n",
- lumv1->lmm_stripe_offset, lov->desc.ld_tgt_count);
- RETURN(-EINVAL);
- }
- stripe_count = lov_get_stripecnt(lov, lmm_magic,
- lumv1->lmm_stripe_count);
-
- if (max_lmm_size) {
- int max_stripes = (max_lmm_size -
- lov_mds_md_size(0, lmm_magic)) /
- sizeof(struct lov_ost_data_v1);
- if (unlikely(max_stripes < stripe_count)) {
- CDEBUG(D_IOCTL, "stripe count reset from %d to %d\n",
- stripe_count, max_stripes);
- stripe_count = max_stripes;
- }
- }
-
- if (lmm_magic == LOV_USER_MAGIC_V3) {
- struct pool_desc *pool;
-
- pool = lov_find_pool(lov, lumv3->lmm_pool_name);
- if (pool != NULL) {
- if (lumv3->lmm_stripe_offset !=
- (typeof(lumv3->lmm_stripe_offset))(-1)) {
- rc = lov_check_index_in_pool(
- lumv3->lmm_stripe_offset, pool);
- if (rc < 0) {
- lov_pool_putref(pool);
- RETURN(-EINVAL);
- }
- }
-
- if (stripe_count > pool_tgt_count(pool))
- stripe_count = pool_tgt_count(pool);
-
- lov_pool_putref(pool);
- }
- }
-
- rc = lov_alloc_memmd(lsmp, stripe_count, lumv1->lmm_pattern, lmm_magic);
-
- if (rc >= 0) {
- (*lsmp)->lsm_oinfo[0]->loi_ost_idx = lumv1->lmm_stripe_offset;
- (*lsmp)->lsm_stripe_size = lumv1->lmm_stripe_size;
- if (lmm_magic == LOV_USER_MAGIC_V3)
- strncpy((*lsmp)->lsm_pool_name, lumv3->lmm_pool_name,
- LOV_MAXPOOLNAME);
- rc = 0;
- }
-
- RETURN(rc);
-}
+ 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, int max_lmm_size,
- struct lov_stripe_md **lsmp, struct lov_user_md *lump)
-{
- int rc;
- mm_segment_t seg;
+ if (buf_size < sizeof(magic))
+ RETURN(ERR_PTR(-EINVAL));
- seg = get_fs();
- set_fs(KERNEL_DS);
+ magic = le32_to_cpu(*(u32 *)buf);
+ op = lsm_op_find(magic);
+ if (!op)
+ RETURN(ERR_PTR(-EINVAL));
- rc = __lov_setstripe(exp, max_lmm_size, lsmp, lump);
- set_fs(seg);
- RETURN(rc);
-}
+ lsm = op->lsm_unpackmd(lov, buf, buf_size);
-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;
- struct lov_user_ost_data_v1 *lmm_objects;
-
- ENTRY;
-
- if (lump->lmm_magic == LOV_USER_MAGIC_V3)
- lmm_objects = ((struct lov_user_md_v3 *)lump)->lmm_objects;
- else
- lmm_objects = lump->lmm_objects;
-
- for (i = 0; i < lump->lmm_stripe_count; i++) {
- __u32 len = sizeof(last_id);
- oexp = lov->lov_tgts[lmm_objects[i].l_ost_idx]->ltd_exp;
- rc = obd_get_info(oexp, sizeof(KEY_LAST_ID), KEY_LAST_ID,
- &len, &last_id, NULL);
- if (rc)
- RETURN(rc);
- if (lmm_objects[i].l_object_id > last_id) {
- CERROR("Setting EA for object > than last id on "
- "ost idx %d "LPD64" > "LPD64" \n",
- lmm_objects[i].l_ost_idx,
- lmm_objects[i].l_object_id, last_id);
- RETURN(-EINVAL);
- }
- }
-
- rc = lov_setstripe(exp, 0, lsmp, lump);
- if (rc)
- RETURN(rc);
-
- for (i = 0; i < lump->lmm_stripe_count; i++) {
- (*lsmp)->lsm_oinfo[i]->loi_ost_idx =
- lmm_objects[i].l_ost_idx;
- (*lsmp)->lsm_oinfo[i]->loi_id = lmm_objects[i].l_object_id;
- (*lsmp)->lsm_oinfo[i]->loi_seq = lmm_objects[i].l_object_seq;
- }
- RETURN(0);
+ RETURN(lsm);
}
-
-/* Retrieve object striping information.
+/*
+ * 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)
{
- /*
- * XXX huge struct allocated on stack.
- */
- /* we use lov_user_md_v3 because it is larger than lov_user_md_v1 */
- struct lov_user_md_v3 lum;
- struct lov_mds_md *lmmk = NULL;
- int rc, lmm_size;
- int lum_size;
- mm_segment_t seg;
- ENTRY;
-
- if (!lsm)
- RETURN(-ENODATA);
-
- /*
- * "Switch to kernel segment" to allow copying from kernel space by
- * copy_{to,from}_user().
- */
- seg = get_fs();
- set_fs(KERNEL_DS);
-
- /* we only need the header part from user space to get lmm_magic and
- * lmm_stripe_count, (the header part is common to v1 and v3) */
- lum_size = sizeof(struct lov_user_md_v1);
- if (cfs_copy_from_user(&lum, lump, lum_size))
- GOTO(out_set, rc = -EFAULT);
- else if ((lum.lmm_magic != LOV_USER_MAGIC) &&
- (lum.lmm_magic != LOV_USER_MAGIC_V3))
- GOTO(out_set, rc = -EINVAL);
-
- if (lum.lmm_stripe_count &&
- (lum.lmm_stripe_count < lsm->lsm_stripe_count)) {
- /* Return right size of stripe to user */
- lum.lmm_stripe_count = lsm->lsm_stripe_count;
- rc = cfs_copy_to_user(lump, &lum, lum_size);
- GOTO(out_set, rc = -EOVERFLOW);
- }
- rc = lov_packmd(exp, &lmmk, lsm);
- if (rc < 0)
- GOTO(out_set, rc);
- lmm_size = rc;
- rc = 0;
-
- /* FIXME: Bug 1185 - copy fields properly when structs change */
- /* struct lov_user_md_v3 and struct lov_mds_md_v3 must be the same */
- CLASSERT(sizeof(lum) == sizeof(struct lov_mds_md_v3));
- CLASSERT(sizeof lum.lmm_objects[0] == sizeof lmmk->lmm_objects[0]);
-
- if ((cpu_to_le32(LOV_MAGIC) != LOV_MAGIC) &&
- ((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);
- }
- if (lum.lmm_magic == LOV_USER_MAGIC) {
- /* User request for v1, we need skip lmm_pool_name */
- if (lmmk->lmm_magic == LOV_MAGIC_V3) {
- memmove((char*)(&lmmk->lmm_stripe_count) +
- sizeof(lmmk->lmm_stripe_count),
- ((struct lov_mds_md_v3*)lmmk)->lmm_objects,
- lmmk->lmm_stripe_count *
- sizeof(struct lov_ost_data_v1));
- lmm_size -= LOV_MAXPOOLNAME;
- }
- } else {
- /* if v3 we just have to update the lum_size */
- lum_size = sizeof(struct lov_user_md_v3);
- }
-
- /* User wasn't expecting this many OST entries */
- if (lum.lmm_stripe_count == 0)
- lmm_size = lum_size;
- else if (lum.lmm_stripe_count < lmmk->lmm_stripe_count)
- GOTO(out_set, rc = -EOVERFLOW);
- /*
- * Have a difference between lov_mds_md & lov_user_md.
- * So we have to re-order the data before copy to user.
- */
- lum.lmm_stripe_count = lmmk->lmm_stripe_count;
- lum.u.lum_layout_gen = lmmk->lmm_layout_gen;
- ((struct lov_user_md*)lmmk)->u.lum_layout_gen = lum.u.lum_layout_gen;
- ((struct lov_user_md*)lmmk)->lmm_stripe_count = lum.lmm_stripe_count;
- if (cfs_copy_to_user(lump, lmmk, lmm_size))
- rc = -EFAULT;
-
- obd_free_diskmd(exp, &lmmk);
-out_set:
- set_fs(seg);
- 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_foreign_md *lfm;
+ struct lov_user_md_v1 lum;
+ size_t lmmk_size, lum_size = 0;
+ ssize_t lmm_size;
+ int rc = 0;
+
+ ENTRY;
+
+ if (lsm->lsm_magic != LOV_MAGIC_V1 && lsm->lsm_magic != LOV_MAGIC_V3 &&
+ lsm->lsm_magic != LOV_MAGIC_COMP_V1 &&
+ lsm->lsm_magic != LOV_MAGIC_FOREIGN) {
+ 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);
+ }
+
+ lmmk_size = lov_comp_md_size(lsm);
+
+ OBD_ALLOC_LARGE(lmmk, lmmk_size);
+ if (!lmmk)
+ 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);
+ } else if (lmmk->lmm_magic == cpu_to_le32(LOV_MAGIC_FOREIGN)) {
+ lfm = (struct lov_foreign_md *)lmmk;
+ __swab32s(&lfm->lfm_magic);
+ __swab32s(&lfm->lfm_length);
+ __swab32s(&lfm->lfm_type);
+ __swab32s(&lfm->lfm_flags);
+ }
+ }
+
+ /*
+ * Legacy appication passes limited buffer, we need to figure out
+ * the user buffer size by the passed in lmm_stripe_count.
+ */
+ if (lsm->lsm_magic != LOV_MAGIC_FOREIGN)
+ 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_lock(cl_obj);
+ cl_object_attr_get(env, cl_obj, &attr);
+ cl_object_attr_unlock(cl_obj);
+
+ /*
+ * 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);
+ lum_size = comp_v1->lcm_entries[i].lcme_size;
+ }
+
+ lmm = comp_md;
+ lmm_size = min(lum_size, lmmk_size);
+ } else {
+ lmm = lmmk;
+ lmm_size = lmmk_size;
+ }
+
+ /**
+ * Return stripe_count=1 instead of 0 for DoM files to avoid
+ * divide-by-zero for older userspace that calls this ioctl,
+ * e.g. lustre ADIO driver.
+ */
+ if ((lum.lmm_stripe_count == 0) && (lum.lmm_pattern & LOV_PATTERN_MDT))
+ lum.lmm_stripe_count = 1;
+ /**
+ * 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);
}