Viewing: qmt_lock.c
// SPDX-License-Identifier: GPL-2.0
/*
* Copyright (c) 2012, 2017, Intel Corporation.
* Use is subject to license terms.
*/
/*
* Author: Johann Lombardi <johann.lombardi@intel.com>
* Author: Niu Yawei <yawei.niu@intel.com>
*/
#define DEBUG_SUBSYSTEM S_LQUOTA
#include <linux/delay.h>
#include <linux/kthread.h>
#include <linux/workqueue.h>
#include <lustre_dlm.h>
#include <lustre_swab.h>
#include <obd_class.h>
#include "qmt_internal.h"
/* intent policy function called from mdt_intent_opc() when the intent is of
* quota type */
int qmt_intent_policy(const struct lu_env *env, struct lu_device *ld,
struct ptlrpc_request *req, struct ldlm_lock **lockp,
int flags)
{
struct qmt_device *qmt = lu2qmt_dev(ld);
struct ldlm_intent *it;
struct quota_body *reqbody;
struct quota_body *repbody;
struct obd_uuid *uuid;
struct lquota_lvb *lvb;
struct ldlm_resource *res = (*lockp)->l_resource;
struct ldlm_reply *ldlm_rep;
int rc, lvb_len;
ENTRY;
req_capsule_extend(&req->rq_pill, &RQF_LDLM_INTENT_QUOTA);
req_capsule_set_size(&req->rq_pill, &RMF_DLM_LVB, RCL_SERVER,
ldlm_lvbo_size(*lockp));
/* extract quota body and intent opc */
it = req_capsule_client_get(&req->rq_pill, &RMF_LDLM_INTENT);
if (it == NULL)
RETURN(err_serious(-EFAULT));
reqbody = req_capsule_client_get(&req->rq_pill, &RMF_QUOTA_BODY);
if (reqbody == NULL)
RETURN(err_serious(-EFAULT));
/* prepare reply */
rc = req_capsule_server_pack(&req->rq_pill);
if (rc != 0) {
CERROR("Can't pack response, rc %d\n", rc);
RETURN(err_serious(rc));
}
repbody = req_capsule_server_get(&req->rq_pill, &RMF_QUOTA_BODY);
if (repbody == NULL)
RETURN(err_serious(-EFAULT));
ldlm_rep = req_capsule_server_get(&req->rq_pill, &RMF_DLM_REP);
if (ldlm_rep == NULL)
RETURN(err_serious(-EFAULT));
uuid = &(*lockp)->l_export->exp_client_uuid;
switch (it->opc) {
case IT_QUOTA_DQACQ: {
struct lquota_entry *lqe;
struct ldlm_lock *lock;
enum qmt_stype stype;
int idx;
if (res->lr_name.name[LUSTRE_RES_ID_QUOTA_SEQ_OFF] == 0)
/* acquire on global lock? something is wrong ... */
GOTO(out, rc = -EPROTO);
/* verify global lock isn't stale */
if (!lustre_handle_is_used(&reqbody->qb_glb_lockh))
GOTO(out, rc = -ENOLCK);
lock = ldlm_handle2lock(&reqbody->qb_glb_lockh);
if (lock == NULL)
GOTO(out, rc = -ENOLCK);
ldlm_lock_put(lock);
stype = qmt_uuid2idx(uuid, &idx);
if (stype < 0)
GOTO(out, rc = -EINVAL);
/* TODO: it seems we don't need to get lqe from
* lq_lvb_data anymore ... And do extra get
* and put on it */
lqe = res->lr_lvb_data;
LASSERT(lqe != NULL);
lqe_getref(lqe);
rc = qmt_pool_lqes_lookup(env, qmt, lqe_rtype(lqe), stype,
lqe_qtype(lqe), &reqbody->qb_id,
NULL, idx);
if (rc) {
lqe_putref(lqe);
GOTO(out, rc);
}
/* acquire quota space */
rc = qmt_dqacq0(env, qmt, uuid,
reqbody->qb_flags, reqbody->qb_count,
reqbody->qb_usage, repbody,
qmt_dom(lqe_rtype(lqe), stype) ? -1 : idx);
lqe_putref(lqe);
qti_lqes_fini(env);
if (rc)
GOTO(out, rc);
break;
}
case IT_QUOTA_CONN:
/* new connection from slave */
if (res->lr_name.name[LUSTRE_RES_ID_QUOTA_SEQ_OFF] != 0)
/* connection on per-ID lock? something is wrong ... */
GOTO(out, rc = -EPROTO);
rc = qmt_pool_new_conn(env, qmt, &reqbody->qb_fid,
&repbody->qb_slv_fid,
&repbody->qb_slv_ver, uuid);
if (rc)
GOTO(out, rc);
break;
default:
CERROR("%s: invalid intent opcode: %llu\n", qmt->qmt_svname,
it->opc);
GOTO(out, rc = -EINVAL);
}
/* on success, pack lvb in reply */
lvb = req_capsule_server_get(&req->rq_pill, &RMF_DLM_LVB);
lvb_len = ldlm_lvbo_size(*lockp);
lvb_len = ldlm_lvbo_fill(*lockp, lvb, &lvb_len);
if (lvb_len < 0)
GOTO(out, rc = lvb_len);
req_capsule_shrink(&req->rq_pill, &RMF_DLM_LVB, lvb_len, RCL_SERVER);
out:
ldlm_rep->lock_policy_res2 = clear_serious(rc);
EXIT;
return ELDLM_OK;
}
/*
* Initialize quota LVB associated with quota indexes.
* Called with res->lr_lvb_sem held
*/
int qmt_lvbo_init(struct lu_device *ld, struct ldlm_resource *res)
{
struct lu_env *env;
struct qmt_thread_info *qti;
struct qmt_device *qmt = lu2qmt_dev(ld);
enum lquota_res_type pool_type;
enum lquota_type qtype;
int rc;
ENTRY;
LASSERT(res != NULL);
if (res->lr_type != LDLM_PLAIN)
RETURN(-ENOTSUPP);
if (res->lr_lvb_data ||
res->lr_name.name[LUSTRE_RES_ID_SEQ_OFF] != FID_SEQ_QUOTA_GLB)
RETURN(0);
env = lu_env_find();
LASSERT(env);
qti = qmt_info(env);
/* extract global index FID and quota identifier */
fid_extract_from_quota_res(&qti->qti_fid, &qti->qti_id, &res->lr_name);
/* sanity check the global index FID */
rc = lquota_extract_fid(&qti->qti_fid, &pool_type, &qtype);
if (rc) {
CERROR("can't extract glb index information from FID "DFID"\n",
PFID(&qti->qti_fid));
GOTO(out, rc);
}
if (res->lr_name.name[LUSTRE_RES_ID_QUOTA_SEQ_OFF] != 0) {
/* no ID quota lock associated with UID/GID 0 or with a seq 0,
* we are thus dealing with an ID lock. */
struct qmt_pool_info *pool;
struct lquota_entry *lqe;
struct lqe_glbl_data *lgd;
int wait = 0;
pool = qmt_pool_lookup_glb(env, qmt, pool_type);
if (IS_ERR(pool))
GOTO(out, rc = -ENOMEM);
/* Find the quota entry associated with the quota id */
lqe = qmt_pool_lqe_lookup(env, qmt, pool_type, qtype,
&qti->qti_id, NULL);
if (IS_ERR(lqe)) {
qpi_putref(env, pool);
GOTO(out, rc = PTR_ERR(lqe));
}
again:
mutex_lock(&lqe->lqe_glbl_data_lock);
/* Is the old lqe_glbl_data still waiting to be freed in
* qmt_lvbo_free_wq?
*/
if (lqe->lqe_glbl_data) {
mutex_unlock(&lqe->lqe_glbl_data_lock);
wait++;
/* wait one second */
if (wait < 1000) {
msleep_interruptible(1);
goto again;
}
LQUOTA_ERROR(lqe, "the lvb is held by qmt_wq: %p\n",
res->lr_lvb_data);
lqe_putref(lqe);
GOTO(out_put_qpi, rc = -EBUSY);
}
/* TODO: need something like qmt_extend_lqe_gd that has
* to be called each time when qpi_slv_nr is incremented
*/
lgd = qmt_alloc_lqe_gd(pool, qtype);
if (!lgd) {
mutex_unlock(&lqe->lqe_glbl_data_lock);
lqe_putref(lqe);
GOTO(out_put_qpi, rc = -ENOMEM);
}
qmt_setup_lqe_gd(env, qmt, lqe, lgd, pool_type);
lqe->lqe_glbl_data = lgd;
mutex_unlock(&lqe->lqe_glbl_data_lock);
qmt_id_lock_notify(qmt, lqe);
/* store reference to lqe in lr_lvb_data */
res->lr_lvb_data = lqe;
LQUOTA_DEBUG(lqe, "initialized res lvb");
out_put_qpi:
qpi_putref(env, pool);
} else {
struct dt_object *obj;
/* lookup global index */
obj = dt_locate(env, qmt->qmt_child, &qti->qti_fid);
if (IS_ERR(obj))
GOTO(out, rc = PTR_ERR(obj));
if (!dt_object_exists(obj)) {
dt_object_put(env, obj);
GOTO(out, rc = -ENOENT);
}
/* store reference to global index object in lr_lvb_data */
res->lr_lvb_data = obj;
CDEBUG(D_QUOTA, DFID" initialized lvb\n", PFID(&qti->qti_fid));
}
res->lr_lvb_len = sizeof(struct lquota_lvb);
EXIT;
out:
return rc;
}
/* clear lge_qunit/edquot_nu flags -
* slave received new qunit and edquot.
*
* \retval true if revoke is needed - qunit
* for this slave reaches least_qunit
*/
static bool qmt_clear_lgeg_arr_nu(struct lquota_entry *lqe, int stype, int idx)
{
unsigned long least = lqe2qpi(lqe)->qpi_least_qunit;
bool revoke = false;
/* There is no array to store lge for the case of DOM.
* Ignore it until MDT pools will be ready.
*/
if (!qmt_dom(lqe_rtype(lqe), stype)) {
struct lqe_glbl_data *lgd;
mutex_lock(&lqe->lqe_glbl_data_lock);
lgd = lqe->lqe_glbl_data;
if (lgd) {
int lge_idx = qmt_map_lge_idx(lgd, idx);
if (lge_idx < 0)
return false;
lgd->lqeg_arr[lge_idx].lge_qunit_nu = 0;
lgd->lqeg_arr[lge_idx].lge_edquot_nu = 0;
/* We shouldn't call revoke for DOM case, it will be
* updated at qmt_id_lock_glimpse.
*/
revoke = lgd->lqeg_arr[lge_idx].lge_qunit == least;
}
mutex_unlock(&lqe->lqe_glbl_data_lock);
}
return revoke;
}
static bool qmt_set_revoke(struct lu_env *env, struct lquota_entry *lqe_gl,
int stype, int idx)
{
unsigned long least_qunit = lqe2qpi(lqe_gl)->qpi_least_qunit;
bool notify = false;
if (qmt_dom(lqe_rtype(lqe_gl), stype))
return false;
qti_lqes_write_lock(env);
mutex_lock(&lqe_gl->lqe_glbl_data_lock);
if (lqe_gl->lqe_glbl_data) {
struct lqe_glbl_data *lgd = lqe_gl->lqe_glbl_data;
int lge_idx;
lge_idx = qmt_map_lge_idx(lgd, idx);
LASSERT(lge_idx >= 0);
if (lgd->lqeg_arr[lge_idx].lge_qunit == least_qunit) {
struct lquota_entry *lqe;
int i;
for (i = 0; i < qti_lqes_cnt(env); i++) {
lqe = qti_lqes(env)[i];
LQUOTA_DEBUG(lqe,
"lge_qunit %llu least_qunit %lu idx %d\n",
lgd->lqeg_arr[lge_idx].lge_qunit,
least_qunit, idx);
if (lqe->lqe_qunit == least_qunit) {
lqe->lqe_revoke_time =
ktime_get_seconds();
notify |= qmt_adjust_edquot(lqe,
ktime_get_real_seconds());
}
}
}
}
mutex_unlock(&lqe_gl->lqe_glbl_data_lock);
qti_lqes_write_unlock(env);
return notify;
}
/*
* Update LVB associated with the global quota index.
* This function is called from the DLM itself after a glimpse callback, in this
* case valid ptlrpc request is passed.
*/
int qmt_lvbo_update(struct lu_device *ld, struct ldlm_resource *res,
struct ptlrpc_request *req, int increase_only)
{
struct lu_env *env;
struct qmt_thread_info *qti;
struct qmt_device *qmt = lu2qmt_dev(ld);
struct lquota_entry *lqe;
struct lquota_lvb *lvb;
struct ldlm_lock *lock;
struct obd_export *exp;
bool need_revoke;
enum qmt_stype stype;
int rc = 0, idx;
ENTRY;
LASSERT(res != NULL);
if (req == NULL)
RETURN(0);
if (res->lr_name.name[LUSTRE_RES_ID_QUOTA_SEQ_OFF] == 0)
/* no need to update lvb for global quota locks */
RETURN(0);
lvb = req_capsule_server_swab_get(&req->rq_pill, &RMF_DLM_LVB,
lustre_swab_lquota_lvb);
if (lvb == NULL) {
CERROR("%s: failed to extract lvb from request\n",
qmt->qmt_svname);
RETURN(-EFAULT);
}
lqe = res->lr_lvb_data;
LASSERT(lqe != NULL);
lqe_getref(lqe);
/* allocate environement */
env = lu_env_find();
LASSERT(env);
qti = qmt_info(env);
/* The request is a glimpse callback which was sent via the
* reverse import to the slave. What we care about here is the
* export associated with the slave and req->rq_export is
* definitely not what we are looking for (it is actually set to
* NULL here).
* Therefore we extract the lock from the request argument
* and use lock->l_export. */
lock = ldlm_request_lock(req);
if (IS_ERR(lock)) {
CERROR("%s: failed to get lock from request!\n",
qmt->qmt_svname);
GOTO(out, rc = PTR_ERR(lock));
}
exp = class_export_get(lock->l_export);
if (exp == NULL) {
CERROR("%s: failed to get export from lock!\n",
qmt->qmt_svname);
GOTO(out, rc = -EFAULT);
}
stype = qmt_uuid2idx(&exp->exp_client_uuid, &idx);
if (stype < 0)
GOTO(out_exp, rc = stype);
need_revoke = qmt_clear_lgeg_arr_nu(lqe, stype, idx);
if (lvb->lvb_id_rel == 0) {
/* nothing to release */
if (lvb->lvb_id_may_rel != 0) {
/* but might still release later ... */
lqe_write_lock(lqe);
lqe->lqe_may_rel += lvb->lvb_id_may_rel;
lqe_write_unlock(lqe);
}
}
if (!need_revoke && lvb->lvb_id_rel == 0)
GOTO(out_exp, rc = 0);
rc = qmt_pool_lqes_lookup(env, qmt, lqe_rtype(lqe), stype,
lqe_qtype(lqe), &lqe->lqe_id, NULL, idx);
if (rc)
GOTO(out_exp, rc);
if (need_revoke && qmt_set_revoke(env, lqe, stype, idx)) {
int notify = false;
mutex_lock(&lqe->lqe_glbl_data_lock);
if (lqe->lqe_glbl_data) {
qmt_seed_glbe_edquot(env, lqe->lqe_glbl_data);
notify = true;
}
mutex_unlock(&lqe->lqe_glbl_data_lock);
if (notify)
qmt_id_lock_notify(qmt, lqe);
}
if (lvb->lvb_id_rel) {
LQUOTA_DEBUG(lqe, "releasing:%llu may release:%llu",
lvb->lvb_id_rel, lvb->lvb_id_may_rel);
/* release quota space */
rc = qmt_dqacq0(env, qmt, &exp->exp_client_uuid,
QUOTA_DQACQ_FL_REL, lvb->lvb_id_rel,
0, &qti->qti_body,
qmt_dom(lqe_rtype(lqe), stype) ? -1 : idx);
if (rc || qti->qti_body.qb_count != lvb->lvb_id_rel)
LQUOTA_ERROR(lqe,
"failed to release quota space on glimpse %llu!=%llu : rc = %d\n",
qti->qti_body.qb_count,
lvb->lvb_id_rel, rc);
}
qti_lqes_fini(env);
if (rc)
GOTO(out_exp, rc);
EXIT;
out_exp:
class_export_put(exp);
out:
lqe_putref(lqe);
return rc;
}
/*
* Report size of lvb to ldlm layer in order to allocate lvb buffer
* As far as quota locks are concerned, the size is static and is the same
* for both global and per-ID locks which shares the same lvb format.
*/
int qmt_lvbo_size(struct lu_device *ld, struct ldlm_lock *lock)
{
return sizeof(struct lquota_lvb);
}
/*
* Fill request buffer with quota lvb
*/
int qmt_lvbo_fill(struct lu_device *ld, struct ldlm_lock *lock, void *lvb,
int lvblen)
{
struct ldlm_resource *res = lock->l_resource;
struct lquota_lvb *qlvb = lvb;
struct lu_env *env;
int rc;
ENTRY;
LASSERT(res != NULL);
rc = 0;
if (res->lr_type != LDLM_PLAIN || res->lr_lvb_data == NULL ||
res->lr_name.name[LUSTRE_RES_ID_SEQ_OFF] != FID_SEQ_QUOTA_GLB)
RETURN(-EINVAL);
env = lu_env_find();
LASSERT(env);
if (res->lr_name.name[LUSTRE_RES_ID_QUOTA_SEQ_OFF] != 0) {
/* no ID quota lock associated with UID/GID 0 or with a seq 0,
* we are thus dealing with an ID lock. */
struct lquota_entry *lqe = res->lr_lvb_data;
struct qmt_device *qmt;
struct obd_uuid *uuid;
enum qmt_stype stype;
int idx;
uuid = &(lock)->l_export->exp_client_uuid;
stype = qmt_uuid2idx(uuid, &idx);
if (stype < 0)
RETURN(stype);
qmt = lu2qmt_dev(ld);
/* return current qunit value & edquot flags in lvb */
lqe_getref(lqe);
rc = qmt_pool_lqes_lookup(env, qmt, lqe_rtype(lqe), stype,
lqe_qtype(lqe), &lqe->lqe_id,
NULL, idx);
if (!rc) {
qlvb->lvb_id_qunit = qti_lqes_min_qunit(env);
qlvb->lvb_flags = 0;
if (qti_lqes_edquot(env))
qlvb->lvb_flags = LQUOTA_FL_EDQUOT;
qti_lqes_fini(env);
}
CDEBUG(D_QUOTA, "uuid %s lqe_id %lu, edquot %llu qunit %llu\n",
(char *)uuid, (unsigned long)lqe->lqe_id.qid_uid,
qlvb->lvb_flags, qlvb->lvb_id_qunit);
lqe_putref(lqe);
} else {
/* global quota lock */
struct dt_object *obj = res->lr_lvb_data;
/* return current version of global index */
qlvb->lvb_glb_ver = dt_version_get(env, obj);
}
RETURN(rc = rc ?: sizeof(struct lquota_lvb));
}
/*
* Free lvb associated with a given ldlm resource
* we don't really allocate a lvb, lr_lvb_data just points to
* the appropriate backend structures.
*/
int qmt_lvbo_free(struct lu_device *ld, struct ldlm_resource *res)
{
struct qmt_device *qmt = lu2qmt_dev(ld);
ENTRY;
if (res->lr_lvb_data == NULL)
RETURN(0);
if (res->lr_name.name[LUSTRE_RES_ID_QUOTA_SEQ_OFF] != 0) {
struct lquota_entry *lqe = res->lr_lvb_data;
queue_work(qmt->qmt_lvbo_free_wq, &lqe->lqe_work);
} else {
struct dt_object *obj = res->lr_lvb_data;
/* release object reference */
dt_object_put(lu_env_find(), obj);
}
res->lr_lvb_data = NULL;
res->lr_lvb_len = 0;
RETURN(0);
}
typedef int (*qmt_glimpse_cb_t)(struct ldlm_lock *, struct lquota_entry *);
struct qmt_gl_lock_array {
unsigned long q_max;
unsigned long q_cnt;
struct ldlm_lock **q_locks;
};
static void qmt_free_lock_array(struct qmt_gl_lock_array *array)
{
int i;
if (array->q_max == 0) {
LASSERT(array->q_locks == NULL);
return;
}
for (i = 0; i < array->q_cnt; i++) {
LASSERT(array->q_locks[i]);
ldlm_lock_put(array->q_locks[i]);
array->q_locks[i] = NULL;
}
array->q_cnt = 0;
OBD_FREE_PTR_ARRAY(array->q_locks, array->q_max);
array->q_locks = NULL;
array->q_max = 0;
}
static int qmt_alloc_lock_array(struct ldlm_resource *res,
struct qmt_gl_lock_array *array,
qmt_glimpse_cb_t cb, void *arg)
{
struct lquota_entry *lqe = arg;
struct list_head *pos;
unsigned long count = 0;
int fail_cnt = 0;
ENTRY;
LASSERT(!array->q_max && !array->q_cnt && !array->q_locks);
again:
if (cb)
mutex_lock(&lqe->lqe_glbl_data_lock);
lock_res(res);
/* scan list of granted locks */
list_for_each(pos, &res->lr_granted) {
struct ldlm_lock *lock;
int rc;
lock = list_entry(pos, struct ldlm_lock, l_res_link);
LASSERT(lock->l_export);
if (cb != NULL) {
rc = cb(lock, arg);
/* slave should not be notified */
if (rc == 0)
continue;
}
count++;
if (array->q_max != 0 && array->q_cnt < array->q_max) {
array->q_locks[array->q_cnt] = ldlm_lock_get(lock);
array->q_cnt++;
}
}
unlock_res(res);
if (cb)
mutex_unlock(&lqe->lqe_glbl_data_lock);
if (count > array->q_max) {
qmt_free_lock_array(array);
if (++fail_cnt > 5)
RETURN(-EAGAIN);
/*
* allocate more slots in case of more qualified locks are
* found during next loop
*/
array->q_max = count + count / 2 + 10;
count = 0;
LASSERT(array->q_locks == NULL && array->q_cnt == 0);
OBD_ALLOC_PTR_ARRAY(array->q_locks, array->q_max);
if (array->q_locks == NULL) {
array->q_max = 0;
RETURN(-ENOMEM);
}
goto again;
}
if (unlikely(count == 0))
qmt_free_lock_array(array);
RETURN(0);
}
static void qmt_setup_id_desc(struct ldlm_lock *lock, union ldlm_gl_desc *desc,
struct lquota_entry *lqe)
{
struct obd_uuid *uuid = &(lock)->l_export->exp_client_uuid;
enum qmt_stype stype;
int idx;
__u64 qunit;
bool edquot;
stype = qmt_uuid2idx(uuid, &idx);
LASSERT(stype >= 0);
/* DOM case - set global lqe settings */
if (qmt_dom(lqe_rtype(lqe), stype)) {
edquot = lqe->lqe_edquot;
qunit = lqe->lqe_qunit;
} else {
struct lqe_glbl_data *lgd;
int lge_idx;
mutex_lock(&lqe->lqe_glbl_data_lock);
lgd = lqe->lqe_glbl_data;
if (lgd) {
lge_idx = qmt_map_lge_idx(lgd, idx);
LASSERT(lge_idx >= 0);
edquot = lgd->lqeg_arr[lge_idx].lge_edquot;
qunit = lgd->lqeg_arr[lge_idx].lge_qunit;
} else {
edquot = lqe->lqe_edquot;
qunit = lqe->lqe_qunit;
}
mutex_unlock(&lqe->lqe_glbl_data_lock);
}
/* fill glimpse descriptor with lqe settings */
desc->lquota_desc.gl_flags = edquot ? LQUOTA_FL_EDQUOT : 0;
desc->lquota_desc.gl_qunit = qunit;
CDEBUG(D_QUOTA, "setup desc: stype %d idx %d, edquot %llu qunit %llu\n",
stype, idx, desc->lquota_desc.gl_flags,
desc->lquota_desc.gl_qunit);
}
/*
* Send glimpse callback to slaves holding a lock on resource \res.
* This is used to notify slaves of new quota settings or to claim quota space
* back.
*
* \param env - is the environment passed by the caller
* \param qmt - is the quota master target
* \param res - is the dlm resource associated with the quota object
* \param desc - is the glimpse descriptor to pack in glimpse callback
* \param cb - is the callback function called on every lock and determine
* whether a glimpse should be issued
* \param arg - is an opaq parameter passed to the callback function
*/
static int qmt_glimpse_lock(const struct lu_env *env, struct qmt_device *qmt,
struct ldlm_resource *res, union ldlm_gl_desc *desc,
qmt_glimpse_cb_t cb, struct lquota_entry *lqe)
{
union ldlm_gl_desc *descs = NULL;
struct list_head *tmp, *pos;
LIST_HEAD(gl_list);
struct qmt_gl_lock_array locks;
unsigned long i, locks_count;
int rc = 0;
ENTRY;
memset(&locks, 0, sizeof(locks));
rc = qmt_alloc_lock_array(res, &locks, cb, lqe);
if (rc) {
CERROR("%s: failed to allocate glimpse lock array (%d)\n",
qmt->qmt_svname, rc);
RETURN(rc);
}
if (!locks.q_cnt) {
CDEBUG(D_QUOTA, "%s: no granted locks to send glimpse\n",
qmt->qmt_svname);
RETURN(0);
}
CDEBUG(D_QUOTA, "found granted locks %lu\n", locks.q_cnt);
locks_count = locks.q_cnt;
/* Use one desc for all works, when called from qmt_glb_lock_notify */
if (cb && locks.q_cnt > 1) {
/* TODO: think about to store this preallocated descs
* in lqe_global in lqeg_arr as a part of lqe_glbl_entry.
* The benefit is that we don't need to allocate/free
* and setup this descs each time. But the drawback is
* memory use (sizeof ldlm_gl_desc * OST_COUNT * user_number).
* for examfple it could be 88 * 256 * 10 000 about 225 MB. */
OBD_ALLOC(descs,
sizeof(struct ldlm_gl_lquota_desc) * locks.q_cnt);
if (!descs) {
CERROR("%s: alloc glimpse lock array failed: rc = %d\n",
qmt->qmt_svname, rc);
qmt_free_lock_array(&locks);
RETURN(-ENOMEM);
}
}
for (i = locks.q_cnt; i > 0; i--) {
struct ldlm_glimpse_work *work;
OBD_ALLOC_PTR(work);
if (work == NULL) {
CERROR("%s: failed to notify a lock.\n",
qmt->qmt_svname);
continue;
}
if (cb) {
if (descs)
desc = &descs[i - 1];
qmt_setup_id_desc(locks.q_locks[i - 1], desc, lqe);
work->gl_interpret_data = lqe;
}
list_add_tail(&work->gl_list, &gl_list);
work->gl_lock = locks.q_locks[i - 1];
work->gl_flags = 0;
work->gl_desc = desc;
locks.q_locks[i - 1] = NULL;
locks.q_cnt--;
}
qmt_free_lock_array(&locks);
if (list_empty(&gl_list)) {
CDEBUG(D_QUOTA, "%s: nobody to notify\n", qmt->qmt_svname);
GOTO(out, rc = 0);
}
/* issue glimpse callbacks to all connected slaves */
rc = ldlm_glimpse_locks(res, &gl_list);
list_for_each_safe(pos, tmp, &gl_list) {
struct ldlm_glimpse_work *work;
work = list_entry(pos, struct ldlm_glimpse_work, gl_list);
list_del(&work->gl_list);
CERROR("%s: failed to notify %s of new quota settings\n",
qmt->qmt_svname,
obd_uuid2str(&work->gl_lock->l_export->exp_client_uuid));
ldlm_lock_put(work->gl_lock);
OBD_FREE_PTR(work);
}
out:
OBD_FREE(descs,
sizeof(struct ldlm_gl_lquota_desc) * locks_count);
RETURN(rc);
}
/*
* Send glimpse request to all global quota locks to push new quota setting to
* slaves.
*
* \param env - is the environment passed by the caller
* \param lqe - is the lquota entry which has new settings
* \param ver - is the version associated with the setting change
*/
void qmt_glb_lock_notify(const struct lu_env *env, struct lquota_entry *lqe,
__u64 ver)
{
struct qmt_thread_info *qti = qmt_info(env);
struct qmt_pool_info *pool = lqe2qpi(lqe);
struct ldlm_resource *res = NULL;
ENTRY;
lquota_generate_fid(&qti->qti_fid, pool->qpi_rtype, lqe_qtype(lqe));
/* send glimpse callback to notify slaves of new quota settings */
qti->qti_gl_desc.lquota_desc.gl_id = lqe->lqe_id;
qti->qti_gl_desc.lquota_desc.gl_flags = 0;
if (lqe->lqe_is_default) {
qti->qti_gl_desc.lquota_desc.gl_hardlimit = 0;
qti->qti_gl_desc.lquota_desc.gl_softlimit = 0;
qti->qti_gl_desc.lquota_desc.gl_time = LQUOTA_GRACE_FLAG(0,
LQUOTA_FLAG_DEFAULT);
} else if (lqe->lqe_is_deleted) {
qti->qti_gl_desc.lquota_desc.gl_hardlimit = 0;
qti->qti_gl_desc.lquota_desc.gl_softlimit = 0;
qti->qti_gl_desc.lquota_desc.gl_time = LQUOTA_GRACE_FLAG(0,
LQUOTA_FLAG_DELETED);
} else if (lqe->lqe_is_reset) {
qti->qti_gl_desc.lquota_desc.gl_hardlimit = lqe->lqe_hardlimit;
qti->qti_gl_desc.lquota_desc.gl_softlimit = lqe->lqe_softlimit;
qti->qti_gl_desc.lquota_desc.gl_time = LQUOTA_GRACE_FLAG(0,
LQUOTA_FLAG_RESET);
} else if (lqe->lqe_granted > lqe->lqe_hardlimit) {
qti->qti_gl_desc.lquota_desc.gl_hardlimit = lqe->lqe_hardlimit;
qti->qti_gl_desc.lquota_desc.gl_softlimit = lqe->lqe_softlimit;
qti->qti_gl_desc.lquota_desc.gl_time = LQUOTA_GRACE_FLAG(0,
LQUOTA_FLAG_REVOKE);
} else {
qti->qti_gl_desc.lquota_desc.gl_hardlimit = lqe->lqe_hardlimit;
qti->qti_gl_desc.lquota_desc.gl_softlimit = lqe->lqe_softlimit;
qti->qti_gl_desc.lquota_desc.gl_time = lqe->lqe_gracetime;
}
qti->qti_gl_desc.lquota_desc.gl_ver = ver;
/* look up ldlm resource associated with global index */
fid_build_reg_res_name(&qti->qti_fid, &qti->qti_resid);
res = ldlm_resource_get(pool->qpi_qmt->qmt_ns, &qti->qti_resid,
LDLM_PLAIN, 0);
if (IS_ERR(res)) {
/* this might happen if no slaves have enqueued global quota
* locks yet */
LQUOTA_DEBUG(lqe, "failed to lookup ldlm resource associated with "
DFID, PFID(&qti->qti_fid));
RETURN_EXIT;
}
qmt_glimpse_lock(env, pool->qpi_qmt, res, &qti->qti_gl_desc,
NULL, NULL);
ldlm_resource_putref(res);
EXIT;
}
/* Callback function used to select locks that should be glimpsed when
* broadcasting the new qunit value */
static int qmt_id_lock_cb(struct ldlm_lock *lock, struct lquota_entry *lqe)
{
struct obd_uuid *uuid = &(lock)->l_export->exp_client_uuid;
struct lqe_glbl_data *lgd = lqe->lqe_glbl_data;
enum qmt_stype stype;
int idx;
stype = qmt_uuid2idx(uuid, &idx);
LASSERT(stype == QMT_STYPE_OST || stype == QMT_STYPE_MDT);
CDEBUG(D_QUOTA, "stype %d rtype %d idx %d uuid %s\n",
stype, lqe_rtype(lqe), idx, uuid->uuid);
/* Quota pools support only OSTs, despite MDTs also could be registered
* as LQUOTA_RES_DT devices(DOM). */
if (qmt_dom(lqe_rtype(lqe), stype))
return 1;
if (lgd) {
int lge_idx = qmt_map_lge_idx(lgd, idx);
LASSERT(lge_idx >= 0);
CDEBUG(D_QUOTA,
"tgt idx:%d lge_idx:%d edquot_nu:%d qunit_nu:%d\n",
idx, lge_idx, lgd->lqeg_arr[lge_idx].lge_edquot_nu,
lgd->lqeg_arr[lge_idx].lge_qunit_nu);
return lgd->lqeg_arr[lge_idx].lge_edquot_nu ||
lgd->lqeg_arr[lge_idx].lge_qunit_nu;
}
return 0;
}
/*
* Send glimpse request on per-ID lock to push new qunit value to slave.
*
* \param env - is the environment passed by the caller
* \param qmt - is the quota master target device
* \param lqe - is the lquota entry with the new qunit value
* \param uuid - is the uuid of the slave acquiring space, if any
*/
static void qmt_id_lock_glimpse(const struct lu_env *env,
struct qmt_device *qmt,
struct lquota_entry *lqe, struct obd_uuid *uuid)
{
struct qmt_thread_info *qti = qmt_info(env);
struct qmt_pool_info *pool = lqe2qpi(lqe);
struct ldlm_resource *res = NULL;
ENTRY;
if (!lqe->lqe_enforced)
RETURN_EXIT;
lquota_generate_fid(&qti->qti_fid, pool->qpi_rtype, lqe_qtype(lqe));
fid_build_quota_res_name(&qti->qti_fid, &lqe->lqe_id, &qti->qti_resid);
res = ldlm_resource_get(qmt->qmt_ns, &qti->qti_resid, LDLM_PLAIN, 0);
if (IS_ERR(res)) {
/* this might legitimately happens if slaves haven't had the
* opportunity to enqueue quota lock yet. */
LQUOTA_DEBUG(lqe, "failed to lookup ldlm resource for per-ID lock "
DFID, PFID(&qti->qti_fid));
lqe_write_lock(lqe);
if (lqe->lqe_revoke_time == 0 &&
lqe->lqe_qunit == pool->qpi_least_qunit)
lqe->lqe_revoke_time = ktime_get_seconds();
lqe_write_unlock(lqe);
RETURN_EXIT;
}
lqe_write_lock(lqe);
/* The purpose of glimpse callback on per-ID lock is twofold:
* - notify slaves of new qunit value and hope they will release some
* spare quota space in return
* - notify slaves that master ran out of quota space and there is no
* need to send acquire request any more until further notice */
/* TODO: it is not clear how to implement below case for all lqes
* from where slaves will be notified in qmt_glimpse_lock. Because
* here we have just global lqe with an array of OSTs that should
* be notified. Theoretically we can find all lqes that includes
* these OSTs, but it is not trivial. So I would propose to move
* this case to another place ... */
if (lqe->lqe_revoke_time == 0 &&
lqe->lqe_qunit == pool->qpi_least_qunit)
/* reset lqe_may_rel, it will be updated on glimpse callback
* replies if needed */
lqe->lqe_may_rel = 0;
/* The rebalance thread is the only thread which can issue glimpses */
LASSERT(!lqe->lqe_gl);
lqe->lqe_gl = true;
lqe_write_unlock(lqe);
/* issue glimpse callback to slaves */
if (lqe->lqe_glbl_data)
qmt_glimpse_lock(env, qmt, res, &qti->qti_gl_desc,
qmt_id_lock_cb, lqe);
lqe_write_lock(lqe);
if (lqe->lqe_revoke_time == 0 &&
lqe->lqe_qunit == pool->qpi_least_qunit) {
lqe->lqe_revoke_time = ktime_get_seconds();
qmt_adjust_edquot(lqe, ktime_get_real_seconds());
}
LASSERT(lqe->lqe_gl);
lqe->lqe_gl = false;
lqe_write_unlock(lqe);
ldlm_resource_putref(res);
EXIT;
}
/*
* Schedule a glimpse request on per-ID locks to push new qunit value or
* edquot flag to quota slaves.
*
* \param qmt - is the quota master target device
* \param lqe - is the lquota entry with the new qunit value
*/
void qmt_id_lock_notify(struct qmt_device *qmt, struct lquota_entry *lqe)
{
bool added = false;
ENTRY;
LASSERT(lqe->lqe_is_global);
lqe_getref(lqe);
spin_lock(&qmt->qmt_reba_lock);
if (!qmt->qmt_stopping && list_empty(&lqe->lqe_link)) {
list_add_tail(&lqe->lqe_link, &qmt->qmt_reba_list);
added = true;
if (qmt->qmt_reba_task)
wake_up_process(qmt->qmt_reba_task);
}
spin_unlock(&qmt->qmt_reba_lock);
if (!added)
lqe_putref(lqe);
EXIT;
}
struct qmt_reba_args {
struct qmt_device *qra_dev;
struct lu_env qra_env;
struct completion *qra_started;
};
#ifndef TASK_IDLE
#define TASK_IDLE TASK_INTERRUPTIBLE
#endif
/*
* The rebalance thread is in charge of sending glimpse callbacks on per-ID
* quota locks owned by slaves in order to notify them of:
* - a qunit shrink in which case slaves might release quota space back in
* glimpse reply.
* - set/clear edquot flag used to cache the "quota exhausted" state of the
* master. When the flag is set, slaves know that there is no need to
* try to acquire quota from the master since this latter has already
* distributed all the space.
*/
static int qmt_reba_thread(void *_args)
{
struct qmt_reba_args *args = _args;
struct qmt_device *qmt = args->qra_dev;
struct lu_env *env = &args->qra_env;
struct lquota_entry *lqe, *tmp;
ENTRY;
complete(args->qra_started);
while (({set_current_state(TASK_IDLE);
!kthread_should_stop(); })) {
spin_lock(&qmt->qmt_reba_lock);
list_for_each_entry_safe(lqe, tmp, &qmt->qmt_reba_list,
lqe_link) {
__set_current_state(TASK_RUNNING);
list_del_init(&lqe->lqe_link);
spin_unlock(&qmt->qmt_reba_lock);
/* lqe_ref == 1 means we hold the last ref,
* so no need to send glimpse callbacks.
*/
if (!kthread_should_stop() &&
kref_read(&lqe->lqe_ref) > 1)
qmt_id_lock_glimpse(env, qmt, lqe, NULL);
lqe_putref(lqe);
spin_lock(&qmt->qmt_reba_lock);
}
spin_unlock(&qmt->qmt_reba_lock);
schedule();
}
__set_current_state(TASK_RUNNING);
lu_env_remove(env);
lu_env_fini(env);
OBD_FREE_PTR(args);
RETURN(0);
}
/*
* Start rebalance thread. Called when the QMT is being setup
*/
int qmt_start_reba_thread(struct qmt_device *qmt)
{
struct task_struct *task;
struct qmt_reba_args *args;
DECLARE_COMPLETION_ONSTACK(started);
int rc;
ENTRY;
OBD_ALLOC_PTR(args);
if (args == NULL)
RETURN(-ENOMEM);
args->qra_dev = qmt;
args->qra_started = &started;
rc = lu_env_init(&args->qra_env, LCT_MD_THREAD);
if (rc) {
CERROR("%s: failed to init env.\n", qmt->qmt_svname);
GOTO(out_env, rc);
}
task = kthread_create(qmt_reba_thread, args,
"qmt_reba_%s", qmt->qmt_svname);
if (IS_ERR(task)) {
CERROR("%s: failed to start rebalance thread (%ld)\n",
qmt->qmt_svname, PTR_ERR(task));
GOTO(out_env_fini, rc = PTR_ERR(task));
}
rc = lu_env_add_task(&args->qra_env, task);
if (rc) {
kthread_stop(task);
GOTO(out_env_fini, rc);
}
qmt->qmt_reba_task = task;
wake_up_process(task);
wait_for_completion(&started);
RETURN(0);
out_env_fini:
lu_env_fini(&args->qra_env);
out_env:
OBD_FREE_PTR(args);
RETURN(rc);
}
/*
* Stop rebalance thread. Called when the QMT is about to shutdown.
*/
void qmt_stop_reba_thread(struct qmt_device *qmt)
{
struct task_struct *task;
spin_lock(&qmt->qmt_reba_lock);
task = qmt->qmt_reba_task;
qmt->qmt_reba_task = NULL;
spin_unlock(&qmt->qmt_reba_lock);
if (task)
kthread_stop(task);
LASSERT(list_empty(&qmt->qmt_reba_list));
}
