X-Git-Url: https://git.whamcloud.com/?a=blobdiff_plain;f=lustre%2Ffid%2Ffid_request.c;h=93f6402a122325dad02077503feaf3ac4a7fe118;hb=f843facff59226d3788d855d1d6948523ab8d944;hp=3d8a40141e7c505e9008f815437ac6ed762cb633;hpb=c0a8fa3749834fa1e6f148ffec8b58ea578c1af4;p=fs%2Flustre-release.git diff --git a/lustre/fid/fid_request.c b/lustre/fid/fid_request.c index 3d8a401..93f6402 100644 --- a/lustre/fid/fid_request.c +++ b/lustre/fid/fid_request.c @@ -15,11 +15,7 @@ * * 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 */ @@ -27,7 +23,7 @@ * Copyright (c) 2007, 2010, Oracle and/or its affiliates. All rights reserved. * Use is subject to license terms. * - * Copyright (c) 2011, 2013, Intel Corporation. + * Copyright (c) 2011, 2017, Intel Corporation. */ /* * This file is part of Lustre, http://www.lustre.org/ @@ -42,8 +38,8 @@ #define DEBUG_SUBSYSTEM S_FID +#include #include -#include #include #include #include @@ -52,6 +48,8 @@ #include #include "fid_internal.h" +struct dentry *seq_debugfs_dir; + static int seq_client_rpc(struct lu_client_seq *seq, struct lu_seq_range *output, __u32 opc, const char *opcname) @@ -76,7 +74,7 @@ static int seq_client_rpc(struct lu_client_seq *seq, /* Zero out input range, this is not recovery yet. */ in = req_capsule_client_get(&req->rq_pill, &RMF_SEQ_RANGE); - range_init(in); + lu_seq_range_init(in); ptlrpc_request_set_replen(req); @@ -110,28 +108,26 @@ static int seq_client_rpc(struct lu_client_seq *seq, debug_mask = D_INFO; } - ptlrpc_at_set_req_timeout(req); + /* Allow seq client RPC during recovery time. */ + req->rq_allow_replay = 1; - if (opc != SEQ_ALLOC_SUPER && seq->lcs_type == LUSTRE_SEQ_METADATA) - mdc_get_rpc_lock(exp->exp_obd->u.cli.cl_rpc_lock, NULL); + ptlrpc_at_set_req_timeout(req); rc = ptlrpc_queue_wait(req); - if (opc != SEQ_ALLOC_SUPER && seq->lcs_type == LUSTRE_SEQ_METADATA) - mdc_put_rpc_lock(exp->exp_obd->u.cli.cl_rpc_lock, NULL); if (rc) GOTO(out_req, rc); out = req_capsule_server_get(&req->rq_pill, &RMF_SEQ_RANGE); *output = *out; - if (!range_is_sane(output)) { + if (!lu_seq_range_is_sane(output)) { CERROR("%s: Invalid range received from server: " DRANGE"\n", seq->lcs_name, PRANGE(output)); GOTO(out_req, rc = -EINVAL); } - if (range_is_exhausted(output)) { + if (lu_seq_range_is_exhausted(output)) { CERROR("%s: Range received from server is exhausted: " DRANGE"]\n", seq->lcs_name, PRANGE(output)); GOTO(out_req, rc = -EINVAL); @@ -200,6 +196,17 @@ static int seq_client_alloc_meta(const struct lu_env *env, * (MDT0)yet */ rc = seq_client_rpc(seq, &seq->lcs_space, SEQ_ALLOC_META, "meta"); + if (rc == -EINPROGRESS || rc == -EAGAIN) { + wait_queue_head_t waitq; + struct l_wait_info lwi; + + /* MDT0 is not ready, let's wait for 2 + * seconds and retry. */ + init_waitqueue_head(&waitq); + lwi = LWI_TIMEOUT(cfs_time_seconds(2), NULL, + NULL); + l_wait_event(waitq, 0, &lwi); + } } while (rc == -EINPROGRESS || rc == -EAGAIN); } @@ -208,18 +215,19 @@ static int seq_client_alloc_meta(const struct lu_env *env, /* Allocate new sequence for client. */ static int seq_client_alloc_seq(const struct lu_env *env, - struct lu_client_seq *seq, seqno_t *seqnr) + struct lu_client_seq *seq, u64 *seqnr) { - int rc; - ENTRY; + int rc; + ENTRY; - LASSERT(range_is_sane(&seq->lcs_space)); + LASSERT(lu_seq_range_is_sane(&seq->lcs_space)); - if (range_is_exhausted(&seq->lcs_space)) { + if (lu_seq_range_is_exhausted(&seq->lcs_space)) { rc = seq_client_alloc_meta(env, seq); if (rc) { - CERROR("%s: Can't allocate new meta-sequence," - "rc %d\n", seq->lcs_name, rc); + if (rc != -EINPROGRESS) + CERROR("%s: Can't allocate new meta-sequence," + "rc = %d\n", seq->lcs_name, rc); RETURN(rc); } else { CDEBUG(D_INFO, "%s: New range - "DRANGE"\n", @@ -229,158 +237,169 @@ static int seq_client_alloc_seq(const struct lu_env *env, rc = 0; } - LASSERT(!range_is_exhausted(&seq->lcs_space)); - *seqnr = seq->lcs_space.lsr_start; - seq->lcs_space.lsr_start += 1; + LASSERT(!lu_seq_range_is_exhausted(&seq->lcs_space)); + *seqnr = seq->lcs_space.lsr_start; + seq->lcs_space.lsr_start += 1; - CDEBUG(D_INFO, "%s: Allocated sequence ["LPX64"]\n", seq->lcs_name, + CDEBUG(D_INFO, "%s: Allocated sequence [%#llx]\n", seq->lcs_name, *seqnr); RETURN(rc); } static int seq_fid_alloc_prep(struct lu_client_seq *seq, - wait_queue_t *link) + wait_queue_entry_t *link) { if (seq->lcs_update) { add_wait_queue(&seq->lcs_waitq, link); set_current_state(TASK_UNINTERRUPTIBLE); mutex_unlock(&seq->lcs_mutex); - waitq_wait(link, TASK_UNINTERRUPTIBLE); + schedule(); mutex_lock(&seq->lcs_mutex); remove_wait_queue(&seq->lcs_waitq, link); set_current_state(TASK_RUNNING); return -EAGAIN; } + ++seq->lcs_update; mutex_unlock(&seq->lcs_mutex); + return 0; } -static void seq_fid_alloc_fini(struct lu_client_seq *seq) +static void seq_fid_alloc_fini(struct lu_client_seq *seq, __u64 seqnr, + bool whole) { LASSERT(seq->lcs_update == 1); + mutex_lock(&seq->lcs_mutex); + if (seqnr != 0) { + CDEBUG(D_INFO, "%s: New sequence [0x%16.16llx]\n", + seq->lcs_name, seqnr); + + seq->lcs_fid.f_seq = seqnr; + if (whole) { + /* Since the caller require the whole seq, + * so marked this seq to be used */ + if (seq->lcs_type == LUSTRE_SEQ_METADATA) + seq->lcs_fid.f_oid = + LUSTRE_METADATA_SEQ_MAX_WIDTH; + else + seq->lcs_fid.f_oid = LUSTRE_DATA_SEQ_MAX_WIDTH; + } else { + seq->lcs_fid.f_oid = LUSTRE_FID_INIT_OID; + } + seq->lcs_fid.f_ver = 0; + } + --seq->lcs_update; - wake_up(&seq->lcs_waitq); + wake_up_all(&seq->lcs_waitq); } /** - * Allocate the whole seq to the caller. - **/ + * Allocate the whole non-used seq to the caller. + * + * \param[in] env pointer to the thread context + * \param[in,out] seq pointer to the client sequence manager + * \param[out] seqnr to hold the new allocated sequence + * + * \retval 0 for new sequence allocated. + * \retval Negative error number on failure. + */ int seq_client_get_seq(const struct lu_env *env, - struct lu_client_seq *seq, seqno_t *seqnr) + struct lu_client_seq *seq, u64 *seqnr) { - wait_queue_t link; + wait_queue_entry_t link; int rc; LASSERT(seqnr != NULL); - mutex_lock(&seq->lcs_mutex); - init_waitqueue_entry_current(&link); - - while (1) { - rc = seq_fid_alloc_prep(seq, &link); - if (rc == 0) - break; - } - - rc = seq_client_alloc_seq(env, seq, seqnr); - if (rc) { - CERROR("%s: Can't allocate new sequence, " - "rc %d\n", seq->lcs_name, rc); - seq_fid_alloc_fini(seq); - mutex_unlock(&seq->lcs_mutex); - return rc; - } - CDEBUG(D_INFO, "%s: allocate sequence " - "[0x%16.16"LPF64"x]\n", seq->lcs_name, *seqnr); + mutex_lock(&seq->lcs_mutex); + init_waitqueue_entry(&link, current); - /* Since the caller require the whole seq, - * so marked this seq to be used */ - if (seq->lcs_type == LUSTRE_SEQ_METADATA) - seq->lcs_fid.f_oid = LUSTRE_METADATA_SEQ_MAX_WIDTH; - else - seq->lcs_fid.f_oid = LUSTRE_DATA_SEQ_MAX_WIDTH; + /* To guarantee that we can get a whole non-used sequence. */ + while (seq_fid_alloc_prep(seq, &link) != 0); - seq->lcs_fid.f_seq = *seqnr; - seq->lcs_fid.f_ver = 0; - /* - * Inform caller that sequence switch is performed to allow it - * to setup FLD for it. - */ - seq_fid_alloc_fini(seq); + rc = seq_client_alloc_seq(env, seq, seqnr); + seq_fid_alloc_fini(seq, rc ? 0 : *seqnr, true); + if (rc) + CERROR("%s: Can't allocate new sequence: rc = %d\n", + seq->lcs_name, rc); mutex_unlock(&seq->lcs_mutex); - return rc; + return rc; } EXPORT_SYMBOL(seq_client_get_seq); -/* Allocate new fid on passed client @seq and save it to @fid. */ +/** + * Allocate new fid on passed client @seq and save it to @fid. + * + * \param[in] env pointer to the thread context + * \param[in,out] seq pointer to the client sequence manager + * \param[out] fid to hold the new allocated fid + * + * \retval 1 for notify the caller that sequence switch + * is performed to allow it to setup FLD for it. + * \retval 0 for new FID allocated in current sequence. + * \retval Negative error number on failure. + */ int seq_client_alloc_fid(const struct lu_env *env, struct lu_client_seq *seq, struct lu_fid *fid) { - wait_queue_t link; + wait_queue_entry_t link; int rc; ENTRY; LASSERT(seq != NULL); LASSERT(fid != NULL); - init_waitqueue_entry_current(&link); + init_waitqueue_entry(&link, current); mutex_lock(&seq->lcs_mutex); if (OBD_FAIL_CHECK(OBD_FAIL_SEQ_EXHAUST)) seq->lcs_fid.f_oid = seq->lcs_width; - while (1) { - seqno_t seqnr; + while (1) { + u64 seqnr; - if (!fid_is_zero(&seq->lcs_fid) && - fid_oid(&seq->lcs_fid) < seq->lcs_width) { - /* Just bump last allocated fid and return to caller. */ - seq->lcs_fid.f_oid += 1; - rc = 0; - break; - } - - rc = seq_fid_alloc_prep(seq, &link); - if (rc) - continue; + if (unlikely(!fid_is_zero(&seq->lcs_fid) && + fid_oid(&seq->lcs_fid) < seq->lcs_width)) { + /* Just bump last allocated fid and return to caller. */ + seq->lcs_fid.f_oid++; + rc = 0; + break; + } - rc = seq_client_alloc_seq(env, seq, &seqnr); - if (rc) { - CERROR("%s: Can't allocate new sequence, " - "rc %d\n", seq->lcs_name, rc); - seq_fid_alloc_fini(seq); + /* Release seq::lcs_mutex via seq_fid_alloc_prep() to avoid + * deadlock during seq_client_alloc_seq(). */ + rc = seq_fid_alloc_prep(seq, &link); + if (rc) + continue; + + rc = seq_client_alloc_seq(env, seq, &seqnr); + /* Re-take seq::lcs_mutex via seq_fid_alloc_fini(). */ + seq_fid_alloc_fini(seq, rc ? 0 : seqnr, false); + if (rc) { + if (rc != -EINPROGRESS) + CERROR("%s: Can't allocate new sequence: " + "rc = %d\n", seq->lcs_name, rc); mutex_unlock(&seq->lcs_mutex); - RETURN(rc); - } - CDEBUG(D_INFO, "%s: Switch to sequence " - "[0x%16.16"LPF64"x]\n", seq->lcs_name, seqnr); - - seq->lcs_fid.f_oid = LUSTRE_FID_INIT_OID; - seq->lcs_fid.f_seq = seqnr; - seq->lcs_fid.f_ver = 0; - - /* - * Inform caller that sequence switch is performed to allow it - * to setup FLD for it. - */ - rc = 1; + RETURN(rc); + } - seq_fid_alloc_fini(seq); - break; - } + rc = 1; + break; + } - *fid = seq->lcs_fid; + *fid = seq->lcs_fid; mutex_unlock(&seq->lcs_mutex); - CDEBUG(D_INFO, "%s: Allocated FID "DFID"\n", seq->lcs_name, PFID(fid)); - RETURN(rc); + CDEBUG(D_INFO, "%s: Allocated FID "DFID"\n", seq->lcs_name, PFID(fid)); + + RETURN(rc); } EXPORT_SYMBOL(seq_client_alloc_fid); @@ -390,10 +409,10 @@ EXPORT_SYMBOL(seq_client_alloc_fid); */ void seq_client_flush(struct lu_client_seq *seq) { - wait_queue_t link; + wait_queue_entry_t link; LASSERT(seq != NULL); - init_waitqueue_entry_current(&link); + init_waitqueue_entry(&link, current); mutex_lock(&seq->lcs_mutex); while (seq->lcs_update) { @@ -401,7 +420,7 @@ void seq_client_flush(struct lu_client_seq *seq) set_current_state(TASK_UNINTERRUPTIBLE); mutex_unlock(&seq->lcs_mutex); - waitq_wait(&link, TASK_UNINTERRUPTIBLE); + schedule(); mutex_lock(&seq->lcs_mutex); remove_wait_queue(&seq->lcs_waitq, &link); @@ -416,58 +435,62 @@ void seq_client_flush(struct lu_client_seq *seq) seq->lcs_space.lsr_index = -1; - range_init(&seq->lcs_space); + lu_seq_range_init(&seq->lcs_space); mutex_unlock(&seq->lcs_mutex); } EXPORT_SYMBOL(seq_client_flush); -static void seq_client_proc_fini(struct lu_client_seq *seq) +static void seq_client_debugfs_fini(struct lu_client_seq *seq) { -#ifdef LPROCFS - ENTRY; - if (seq->lcs_proc_dir) { - if (!IS_ERR(seq->lcs_proc_dir)) - lprocfs_remove(&seq->lcs_proc_dir); - seq->lcs_proc_dir = NULL; - } - EXIT; -#endif /* LPROCFS */ + if (!IS_ERR_OR_NULL(seq->lcs_debugfs_entry)) + ldebugfs_remove(&seq->lcs_debugfs_entry); } -static int seq_client_proc_init(struct lu_client_seq *seq) +static int seq_client_debugfs_init(struct lu_client_seq *seq) { -#ifdef LPROCFS int rc; - ENTRY; - seq->lcs_proc_dir = lprocfs_seq_register(seq->lcs_name, - seq_type_proc_dir, - NULL, NULL); - if (IS_ERR(seq->lcs_proc_dir)) { - CERROR("%s: LProcFS failed in seq-init\n", - seq->lcs_name); - rc = PTR_ERR(seq->lcs_proc_dir); - RETURN(rc); + seq->lcs_debugfs_entry = ldebugfs_register(seq->lcs_name, + seq_debugfs_dir, + NULL, NULL); + if (IS_ERR_OR_NULL(seq->lcs_debugfs_entry)) { + CERROR("%s: LdebugFS failed in seq-init\n", seq->lcs_name); + rc = seq->lcs_debugfs_entry ? PTR_ERR(seq->lcs_debugfs_entry) + : -ENOMEM; + seq->lcs_debugfs_entry = NULL; + RETURN(rc); } - rc = lprocfs_seq_add_vars(seq->lcs_proc_dir, - seq_client_proc_list, seq); - if (rc) { - CERROR("%s: Can't init sequence manager " - "proc, rc %d\n", seq->lcs_name, rc); + rc = ldebugfs_add_vars(seq->lcs_debugfs_entry, + seq_client_debugfs_list, seq); + if (rc) { + CERROR("%s: Can't init sequence manager debugfs, rc %d\n", + seq->lcs_name, rc); GOTO(out_cleanup, rc); } RETURN(0); out_cleanup: - seq_client_proc_fini(seq); + seq_client_debugfs_fini(seq); return rc; +} -#else /* LPROCFS */ - return 0; -#endif +void seq_client_fini(struct lu_client_seq *seq) +{ + ENTRY; + + seq_client_debugfs_fini(seq); + + if (seq->lcs_exp != NULL) { + class_export_put(seq->lcs_exp); + seq->lcs_exp = NULL; + } + + seq->lcs_srv = NULL; + EXIT; } +EXPORT_SYMBOL(seq_client_fini); int seq_client_init(struct lu_client_seq *seq, struct obd_export *exp, @@ -500,29 +523,13 @@ int seq_client_init(struct lu_client_seq *seq, snprintf(seq->lcs_name, sizeof(seq->lcs_name), "cli-%s", prefix); - rc = seq_client_proc_init(seq); + rc = seq_client_debugfs_init(seq); if (rc) seq_client_fini(seq); RETURN(rc); } EXPORT_SYMBOL(seq_client_init); -void seq_client_fini(struct lu_client_seq *seq) -{ - ENTRY; - - seq_client_proc_fini(seq); - - if (seq->lcs_exp != NULL) { - class_export_put(seq->lcs_exp); - seq->lcs_exp = NULL; - } - - seq->lcs_srv = NULL; - EXIT; -} -EXPORT_SYMBOL(seq_client_fini); - int client_fid_init(struct obd_device *obd, struct obd_export *exp, enum lu_cli_type type) { @@ -531,26 +538,30 @@ int client_fid_init(struct obd_device *obd, int rc; ENTRY; + down_write(&cli->cl_seq_rwsem); OBD_ALLOC_PTR(cli->cl_seq); - if (cli->cl_seq == NULL) - RETURN(-ENOMEM); + if (!cli->cl_seq) + GOTO(out, rc = -ENOMEM); OBD_ALLOC(prefix, MAX_OBD_NAME + 5); - if (prefix == NULL) - GOTO(out_free_seq, rc = -ENOMEM); + if (!prefix) + GOTO(out, rc = -ENOMEM); snprintf(prefix, MAX_OBD_NAME + 5, "cli-%s", obd->obd_name); /* Init client side sequence-manager */ rc = seq_client_init(cli->cl_seq, exp, type, prefix, NULL); OBD_FREE(prefix, MAX_OBD_NAME + 5); - if (rc) - GOTO(out_free_seq, rc); - RETURN(rc); -out_free_seq: - OBD_FREE_PTR(cli->cl_seq); - cli->cl_seq = NULL; + GOTO(out, rc); + +out: + if (rc && cli->cl_seq) { + OBD_FREE_PTR(cli->cl_seq); + cli->cl_seq = NULL; + } + up_write(&cli->cl_seq_rwsem); + return rc; } EXPORT_SYMBOL(client_fid_init); @@ -560,47 +571,45 @@ int client_fid_fini(struct obd_device *obd) struct client_obd *cli = &obd->u.cli; ENTRY; - if (cli->cl_seq != NULL) { + down_write(&cli->cl_seq_rwsem); + if (cli->cl_seq) { seq_client_fini(cli->cl_seq); OBD_FREE_PTR(cli->cl_seq); cli->cl_seq = NULL; } + up_write(&cli->cl_seq_rwsem); RETURN(0); } EXPORT_SYMBOL(client_fid_fini); -struct proc_dir_entry *seq_type_proc_dir; - -static int __init fid_mod_init(void) +static int __init fid_init(void) { - seq_type_proc_dir = lprocfs_seq_register(LUSTRE_SEQ_NAME, - proc_lustre_root, - NULL, NULL); - if (IS_ERR(seq_type_proc_dir)) - return PTR_ERR(seq_type_proc_dir); - -# ifdef HAVE_SERVER_SUPPORT - fid_server_mod_init(); -# endif +#ifdef HAVE_SERVER_SUPPORT + int rc = fid_server_mod_init(); - return 0; + if (rc) + return rc; +#endif + seq_debugfs_dir = ldebugfs_register(LUSTRE_SEQ_NAME, + debugfs_lustre_root, + NULL, NULL); + return PTR_ERR_OR_ZERO(seq_debugfs_dir); } -static void __exit fid_mod_exit(void) +static void __exit fid_exit(void) { # ifdef HAVE_SERVER_SUPPORT fid_server_mod_exit(); # endif - - if (seq_type_proc_dir != NULL && !IS_ERR(seq_type_proc_dir)) { - lprocfs_remove(&seq_type_proc_dir); - seq_type_proc_dir = NULL; - } + if (!IS_ERR_OR_NULL(seq_debugfs_dir)) + ldebugfs_remove(&seq_debugfs_dir); } -MODULE_AUTHOR("Sun Microsystems, Inc. "); -MODULE_DESCRIPTION("Lustre FID Module"); +MODULE_AUTHOR("OpenSFS, Inc. "); +MODULE_DESCRIPTION("Lustre File IDentifier"); +MODULE_VERSION(LUSTRE_VERSION_STRING); MODULE_LICENSE("GPL"); -cfs_module(fid, "0.1.0", fid_mod_init, fid_mod_exit); +module_init(fid_init); +module_exit(fid_exit);