static int osc_io_ladvise_start(const struct lu_env *env,
const struct cl_io_slice *slice)
{
int result = 0;
struct cl_io *io = slice->cis_io;
struct osc_io *oio = cl2osc_io(env, slice);
struct cl_object *obj = slice->cis_obj;
struct lov_oinfo *loi = cl2osc(obj)->oo_oinfo;
struct cl_ladvise_io *lio = &io->u.ci_ladvise;
struct obdo *oa = &oio->oi_oa;
struct osc_async_cbargs *cbargs = &oio->oi_cbarg;
struct lu_ladvise *ladvise;
struct ladvise_hdr *ladvise_hdr;
int buf_size;
int num_advise = 1;
ENTRY;
/* TODO: add multiple ladvise support in CLIO */
buf_size = offsetof(typeof(*ladvise_hdr), lah_advise[num_advise]);
if (osc_env_info(env)->oti_ladvise_buf.lb_len < buf_size)
lu_buf_realloc(&osc_env_info(env)->oti_ladvise_buf, buf_size);
ladvise_hdr = osc_env_info(env)->oti_ladvise_buf.lb_buf;
if (ladvise_hdr == NULL)
RETURN(-ENOMEM);
memset(ladvise_hdr, 0, buf_size);
ladvise_hdr->lah_magic = LADVISE_MAGIC;
ladvise_hdr->lah_count = num_advise;
ladvise_hdr->lah_flags = lio->li_flags;
memset(oa, 0, sizeof(*oa));
oa->o_oi = loi->loi_oi;
oa->o_valid = OBD_MD_FLID;
obdo_set_parent_fid(oa, lio->li_fid);
ladvise = ladvise_hdr->lah_advise;
ladvise->lla_start = lio->li_start;
ladvise->lla_end = lio->li_end;
ladvise->lla_advice = lio->li_advice;
if (lio->li_flags & LF_ASYNC) {
result = osc_ladvise_base(osc_export(cl2osc(obj)), oa,
ladvise_hdr, NULL, NULL, NULL);
} else {
init_completion(&cbargs->opc_sync);
result = osc_ladvise_base(osc_export(cl2osc(obj)), oa,
ladvise_hdr, osc_async_upcall,
cbargs, PTLRPCD_SET);
cbargs->opc_rpc_sent = result == 0;
}
RETURN(result);
}
static int osc_fsync_ost(const struct lu_env *env, struct osc_object *obj,
struct cl_fsync_io *fio)
{
struct osc_io *oio = osc_env_io(env);
struct obdo *oa = &oio->oi_oa;
struct obd_info *oinfo = &oio->oi_info;
struct lov_oinfo *loi = obj->oo_oinfo;
struct osc_async_cbargs *cbargs = &oio->oi_cbarg;
int rc = 0;
memset(oa, 0, sizeof(*oa));
oa->o_oi = loi->loi_oi;
oa->o_valid = OBD_MD_FLID | OBD_MD_FLGROUP;
/* reload size abd blocks for start and end of sync range */
oa->o_size = fio->fi_start;
oa->o_blocks = fio->fi_end;
oa->o_valid |= OBD_MD_FLSIZE | OBD_MD_FLBLOCKS;
obdo_set_parent_fid(oa, fio->fi_fid);
memset(oinfo, 0, sizeof(*oinfo));
oinfo->oi_oa = oa;
oinfo->oi_capa = fio->fi_capa;
init_completion(&cbargs->opc_sync);
rc = osc_sync_base(osc_export(obj), oinfo, osc_async_upcall, cbargs,
PTLRPCD_SET);
return rc;
}
static int osc_io_commit_write(const struct lu_env *env,
const struct cl_io_slice *ios,
const struct cl_page_slice *slice,
unsigned from, unsigned to)
{
struct osc_io *oio = cl2osc_io(env, ios);
struct osc_page *opg = cl2osc_page(slice);
struct osc_object *obj = cl2osc(opg->ops_cl.cpl_obj);
struct osc_async_page *oap = &opg->ops_oap;
LASSERT(to > 0);
/*
* XXX instead of calling osc_page_touch() here and in
* osc_io_fault_start() it might be more logical to introduce
* cl_page_touch() method, that generic cl_io_commit_write() and page
* fault code calls.
*/
osc_page_touch(env, cl2osc_page(slice), to);
if (!client_is_remote(osc_export(obj)) &&
capable(CFS_CAP_SYS_RESOURCE))
oap->oap_brw_flags |= OBD_BRW_NOQUOTA;
if (oio->oi_lockless)
/* see osc_io_prepare_write() for lockless io handling. */
cl_page_clip(env, slice->cpl_page, from, to);
return 0;
}
static int osc_object_prune(const struct lu_env *env, struct cl_object *obj)
{
struct osc_object *osc = cl2osc(obj);
struct ldlm_res_id *resname = &osc_env_info(env)->oti_resname;
/* DLM locks don't hold a reference of osc_object so we have to
* clear it before the object is being destroyed. */
ostid_build_res_name(&osc->oo_oinfo->loi_oi, resname);
ldlm_resource_iterate(osc_export(osc)->exp_obd->obd_namespace, resname,
osc_object_ast_clear, osc);
return 0;
}
static int osc_io_data_version_start(const struct lu_env *env,
const struct cl_io_slice *slice)
{
struct cl_data_version_io *dv = &slice->cis_io->u.ci_data_version;
struct osc_io *oio = cl2osc_io(env, slice);
struct obdo *oa = &oio->oi_oa;
struct osc_async_cbargs *cbargs = &oio->oi_cbarg;
struct osc_object *obj = cl2osc(slice->cis_obj);
struct lov_oinfo *loi = obj->oo_oinfo;
struct obd_export *exp = osc_export(obj);
struct ptlrpc_request *req;
struct ost_body *body;
struct osc_data_version_args *dva;
int rc;
ENTRY;
memset(oa, 0, sizeof(*oa));
oa->o_oi = loi->loi_oi;
oa->o_valid = OBD_MD_FLID | OBD_MD_FLGROUP;
if (dv->dv_flags & (LL_DV_RD_FLUSH | LL_DV_WR_FLUSH)) {
oa->o_valid |= OBD_MD_FLFLAGS;
oa->o_flags |= OBD_FL_SRVLOCK;
if (dv->dv_flags & LL_DV_WR_FLUSH)
oa->o_flags |= OBD_FL_FLUSH;
}
init_completion(&cbargs->opc_sync);
req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_GETATTR);
if (req == NULL)
RETURN(-ENOMEM);
rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_GETATTR);
if (rc < 0) {
ptlrpc_request_free(req);
RETURN(rc);
}
body = req_capsule_client_get(&req->rq_pill, &RMF_OST_BODY);
lustre_set_wire_obdo(&req->rq_import->imp_connect_data, &body->oa, oa);
ptlrpc_request_set_replen(req);
req->rq_interpret_reply = osc_data_version_interpret;
CLASSERT(sizeof(*dva) <= sizeof(req->rq_async_args));
dva = ptlrpc_req_async_args(req);
dva->dva_oio = oio;
ptlrpcd_add_req(req);
RETURN(0);
}
static int osc_object_prune(const struct lu_env *env, struct cl_object *obj)
{
struct osc_object *osc = cl2osc(obj);
struct ldlm_res_id *resname = &osc_env_info(env)->oti_resname;
LASSERTF(osc->oo_npages == 0,
DFID "still have %lu pages, obj: %p, osc: %p\n",
PFID(lu_object_fid(&obj->co_lu)), osc->oo_npages, obj, osc);
/* DLM locks don't hold a reference of osc_object so we have to
* clear it before the object is being destroyed. */
ostid_build_res_name(&osc->oo_oinfo->loi_oi, resname);
ldlm_resource_iterate(osc_export(osc)->exp_obd->obd_namespace, resname,
osc_object_ast_clear, osc);
return 0;
}
/**
* Finds an existing lock covering a page with given index.
* Copy of osc_obj_dlmlock_at_pgoff() but for DoM IBITS lock.
*/
struct ldlm_lock *mdc_dlmlock_at_pgoff(const struct lu_env *env,
struct osc_object *obj, pgoff_t index,
enum osc_dap_flags dap_flags)
{
struct osc_thread_info *info = osc_env_info(env);
struct ldlm_res_id *resname = &info->oti_resname;
union ldlm_policy_data *policy = &info->oti_policy;
struct lustre_handle lockh;
struct ldlm_lock *lock = NULL;
enum ldlm_mode mode;
__u64 flags;
ENTRY;
fid_build_reg_res_name(lu_object_fid(osc2lu(obj)), resname);
mdc_lock_build_policy(env, policy);
flags = LDLM_FL_BLOCK_GRANTED | LDLM_FL_CBPENDING;
if (dap_flags & OSC_DAP_FL_TEST_LOCK)
flags |= LDLM_FL_TEST_LOCK;
again:
/* Next, search for already existing extent locks that will cover us */
/* If we're trying to read, we also search for an existing PW lock. The
* VFS and page cache already protect us locally, so lots of readers/
* writers can share a single PW lock. */
mode = mdc_dom_lock_match(env, osc_export(obj), resname, LDLM_IBITS,
policy, LCK_PR | LCK_PW, &flags, obj, &lockh,
dap_flags & OSC_DAP_FL_CANCELING);
if (mode != 0) {
lock = ldlm_handle2lock(&lockh);
/* RACE: the lock is cancelled so let's try again */
if (unlikely(lock == NULL))
goto again;
}
RETURN(lock);
}
/**
* Implementation of struct cl_object_operations::coo_req_attr_set() for osc
* layer. osc is responsible for struct obdo::o_id and struct obdo::o_seq
* fields.
*/
static void osc_req_attr_set(const struct lu_env *env, struct cl_object *obj,
struct cl_req_attr *attr)
{
struct lov_oinfo *oinfo;
struct obdo *oa;
struct ost_lvb *lvb;
u64 flags = attr->cra_flags;
oinfo = cl2osc(obj)->oo_oinfo;
lvb = &oinfo->loi_lvb;
oa = attr->cra_oa;
if ((flags & OBD_MD_FLMTIME) != 0) {
oa->o_mtime = lvb->lvb_mtime;
oa->o_valid |= OBD_MD_FLMTIME;
}
if ((flags & OBD_MD_FLATIME) != 0) {
oa->o_atime = lvb->lvb_atime;
oa->o_valid |= OBD_MD_FLATIME;
}
if ((flags & OBD_MD_FLCTIME) != 0) {
oa->o_ctime = lvb->lvb_ctime;
oa->o_valid |= OBD_MD_FLCTIME;
}
if (flags & OBD_MD_FLGROUP) {
ostid_set_seq(&oa->o_oi, ostid_seq(&oinfo->loi_oi));
oa->o_valid |= OBD_MD_FLGROUP;
}
if (flags & OBD_MD_FLID) {
ostid_set_id(&oa->o_oi, ostid_id(&oinfo->loi_oi));
oa->o_valid |= OBD_MD_FLID;
}
if (flags & OBD_MD_FLHANDLE) {
struct ldlm_lock *lock;
struct osc_page *opg;
opg = osc_cl_page_osc(attr->cra_page, cl2osc(obj));
lock = osc_dlmlock_at_pgoff(env, cl2osc(obj), osc_index(opg),
OSC_DAP_FL_TEST_LOCK | OSC_DAP_FL_CANCELING);
if (lock == NULL && !opg->ops_srvlock) {
struct ldlm_resource *res;
struct ldlm_res_id *resname;
CL_PAGE_DEBUG(D_ERROR, env, attr->cra_page,
"uncovered page!\n");
resname = &osc_env_info(env)->oti_resname;
ostid_build_res_name(&oinfo->loi_oi, resname);
res = ldlm_resource_get(
osc_export(cl2osc(obj))->exp_obd->obd_namespace,
NULL, resname, LDLM_EXTENT, 0);
ldlm_resource_dump(D_ERROR, res);
libcfs_debug_dumpstack(NULL);
LBUG();
}
/* check for lockless io. */
if (lock != NULL) {
oa->o_handle = lock->l_remote_handle;
oa->o_valid |= OBD_MD_FLHANDLE;
LDLM_LOCK_PUT(lock);
}
}
}
static int osc_object_fiemap(const struct lu_env *env, struct cl_object *obj,
struct ll_fiemap_info_key *fmkey,
struct fiemap *fiemap, size_t *buflen)
{
struct obd_export *exp = osc_export(cl2osc(obj));
struct ldlm_res_id resid;
union ldlm_policy_data policy;
struct lustre_handle lockh;
enum ldlm_mode mode = LCK_MINMODE;
struct ptlrpc_request *req;
struct fiemap *reply;
char *tmp;
int rc;
ENTRY;
fmkey->lfik_oa.o_oi = cl2osc(obj)->oo_oinfo->loi_oi;
if (!(fmkey->lfik_fiemap.fm_flags & FIEMAP_FLAG_SYNC))
goto skip_locking;
policy.l_extent.start = fmkey->lfik_fiemap.fm_start & PAGE_MASK;
if (OBD_OBJECT_EOF - fmkey->lfik_fiemap.fm_length <=
fmkey->lfik_fiemap.fm_start + PAGE_SIZE - 1)
policy.l_extent.end = OBD_OBJECT_EOF;
else
policy.l_extent.end = (fmkey->lfik_fiemap.fm_start +
fmkey->lfik_fiemap.fm_length +
PAGE_SIZE - 1) & PAGE_MASK;
ostid_build_res_name(&fmkey->lfik_oa.o_oi, &resid);
mode = ldlm_lock_match(exp->exp_obd->obd_namespace,
LDLM_FL_BLOCK_GRANTED | LDLM_FL_LVB_READY,
&resid, LDLM_EXTENT, &policy,
LCK_PR | LCK_PW, &lockh, 0);
if (mode) { /* lock is cached on client */
if (mode != LCK_PR) {
ldlm_lock_addref(&lockh, LCK_PR);
ldlm_lock_decref(&lockh, LCK_PW);
}
} else { /* no cached lock, needs acquire lock on server side */
fmkey->lfik_oa.o_valid |= OBD_MD_FLFLAGS;
fmkey->lfik_oa.o_flags |= OBD_FL_SRVLOCK;
}
skip_locking:
req = ptlrpc_request_alloc(class_exp2cliimp(exp),
&RQF_OST_GET_INFO_FIEMAP);
if (req == NULL)
GOTO(drop_lock, rc = -ENOMEM);
req_capsule_set_size(&req->rq_pill, &RMF_FIEMAP_KEY, RCL_CLIENT,
sizeof(*fmkey));
req_capsule_set_size(&req->rq_pill, &RMF_FIEMAP_VAL, RCL_CLIENT,
*buflen);
req_capsule_set_size(&req->rq_pill, &RMF_FIEMAP_VAL, RCL_SERVER,
*buflen);
rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_GET_INFO);
if (rc != 0) {
ptlrpc_request_free(req);
GOTO(drop_lock, rc);
}
tmp = req_capsule_client_get(&req->rq_pill, &RMF_FIEMAP_KEY);
memcpy(tmp, fmkey, sizeof(*fmkey));
tmp = req_capsule_client_get(&req->rq_pill, &RMF_FIEMAP_VAL);
memcpy(tmp, fiemap, *buflen);
ptlrpc_request_set_replen(req);
rc = ptlrpc_queue_wait(req);
if (rc != 0)
GOTO(fini_req, rc);
reply = req_capsule_server_get(&req->rq_pill, &RMF_FIEMAP_VAL);
if (reply == NULL)
GOTO(fini_req, rc = -EPROTO);
memcpy(fiemap, reply, *buflen);
fini_req:
ptlrpc_req_finished(req);
drop_lock:
if (mode)
ldlm_lock_decref(&lockh, LCK_PR);
RETURN(rc);
}
static int osc_io_setattr_start(const struct lu_env *env,
const struct cl_io_slice *slice)
{
struct cl_io *io = slice->cis_io;
struct osc_io *oio = cl2osc_io(env, slice);
struct cl_object *obj = slice->cis_obj;
struct lov_oinfo *loi = cl2osc(obj)->oo_oinfo;
struct cl_attr *attr = &osc_env_info(env)->oti_attr;
struct obdo *oa = &oio->oi_oa;
struct osc_async_cbargs *cbargs = &oio->oi_cbarg;
__u64 size = io->u.ci_setattr.sa_attr.lvb_size;
unsigned int ia_valid = io->u.ci_setattr.sa_valid;
int result = 0;
struct obd_info oinfo = { { { 0 } } };
/* truncate cache dirty pages first */
if (cl_io_is_trunc(io))
result = osc_cache_truncate_start(env, oio, cl2osc(obj), size);
if (result == 0 && oio->oi_lockless == 0) {
cl_object_attr_lock(obj);
result = cl_object_attr_get(env, obj, attr);
if (result == 0) {
struct ost_lvb *lvb = &io->u.ci_setattr.sa_attr;
unsigned int cl_valid = 0;
if (ia_valid & ATTR_SIZE) {
attr->cat_size = attr->cat_kms = size;
cl_valid = (CAT_SIZE | CAT_KMS);
}
if (ia_valid & ATTR_MTIME_SET) {
attr->cat_mtime = lvb->lvb_mtime;
cl_valid |= CAT_MTIME;
}
if (ia_valid & ATTR_ATIME_SET) {
attr->cat_atime = lvb->lvb_atime;
cl_valid |= CAT_ATIME;
}
if (ia_valid & ATTR_CTIME_SET) {
attr->cat_ctime = lvb->lvb_ctime;
cl_valid |= CAT_CTIME;
}
result = cl_object_attr_set(env, obj, attr, cl_valid);
}
cl_object_attr_unlock(obj);
}
memset(oa, 0, sizeof(*oa));
if (result == 0) {
oa->o_oi = loi->loi_oi;
oa->o_mtime = attr->cat_mtime;
oa->o_atime = attr->cat_atime;
oa->o_ctime = attr->cat_ctime;
oa->o_valid = OBD_MD_FLID | OBD_MD_FLGROUP | OBD_MD_FLATIME |
OBD_MD_FLCTIME | OBD_MD_FLMTIME;
if (ia_valid & ATTR_SIZE) {
oa->o_size = size;
oa->o_blocks = OBD_OBJECT_EOF;
oa->o_valid |= OBD_MD_FLSIZE | OBD_MD_FLBLOCKS;
if (oio->oi_lockless) {
oa->o_flags = OBD_FL_SRVLOCK;
oa->o_valid |= OBD_MD_FLFLAGS;
}
} else {
LASSERT(oio->oi_lockless == 0);
}
oinfo.oi_oa = oa;
oinfo.oi_capa = io->u.ci_setattr.sa_capa;
init_completion(&cbargs->opc_sync);
if (ia_valid & ATTR_SIZE)
result = osc_punch_base(osc_export(cl2osc(obj)),
&oinfo, osc_async_upcall,
cbargs, PTLRPCD_SET);
else
result = osc_setattr_async_base(osc_export(cl2osc(obj)),
&oinfo, NULL,
osc_async_upcall,
cbargs, PTLRPCD_SET);
cbargs->opc_rpc_sent = result == 0;
}
return result;
}