void obdo_from_iattr(struct obdo *oa, struct iattr *attr, unsigned int ia_valid)
{
if (ia_valid & ATTR_ATIME) {
oa->o_atime = LTIME_S(attr->ia_atime);
oa->o_valid |= OBD_MD_FLATIME;
}
if (ia_valid & ATTR_MTIME) {
oa->o_mtime = LTIME_S(attr->ia_mtime);
oa->o_valid |= OBD_MD_FLMTIME;
}
if (ia_valid & ATTR_CTIME) {
oa->o_ctime = LTIME_S(attr->ia_ctime);
oa->o_valid |= OBD_MD_FLCTIME;
}
if (ia_valid & ATTR_SIZE) {
oa->o_size = attr->ia_size;
oa->o_valid |= OBD_MD_FLSIZE;
}
if (ia_valid & ATTR_MODE) {
oa->o_mode = attr->ia_mode;
oa->o_valid |= OBD_MD_FLTYPE | OBD_MD_FLMODE;
if (!in_group_p(make_kgid(&init_user_ns, oa->o_gid)) &&
!cfs_capable(CFS_CAP_FSETID))
oa->o_mode &= ~S_ISGID;
}
if (ia_valid & ATTR_UID) {
oa->o_uid = from_kuid(&init_user_ns, attr->ia_uid);
oa->o_valid |= OBD_MD_FLUID;
}
if (ia_valid & ATTR_GID) {
oa->o_gid = from_kgid(&init_user_ns, attr->ia_gid);
oa->o_valid |= OBD_MD_FLGID;
}
}
static void mdc_setattr_pack_rec(struct mdt_rec_setattr *rec,
struct md_op_data *op_data)
{
rec->sa_opcode = REINT_SETATTR;
rec->sa_fsuid = from_kuid(&init_user_ns, current_fsuid());
rec->sa_fsgid = from_kgid(&init_user_ns, current_fsgid());
rec->sa_cap = cfs_curproc_cap_pack();
rec->sa_suppgid = -1;
rec->sa_fid = op_data->op_fid1;
rec->sa_valid = attr_pack(op_data->op_attr.ia_valid);
rec->sa_mode = op_data->op_attr.ia_mode;
rec->sa_uid = from_kuid(&init_user_ns, op_data->op_attr.ia_uid);
rec->sa_gid = from_kgid(&init_user_ns, op_data->op_attr.ia_gid);
rec->sa_size = op_data->op_attr.ia_size;
rec->sa_blocks = op_data->op_attr_blocks;
rec->sa_atime = LTIME_S(op_data->op_attr.ia_atime);
rec->sa_mtime = LTIME_S(op_data->op_attr.ia_mtime);
rec->sa_ctime = LTIME_S(op_data->op_attr.ia_ctime);
rec->sa_attr_flags = ((struct ll_iattr *)&op_data->op_attr)->ia_attr_flags;
if ((op_data->op_attr.ia_valid & ATTR_GID) &&
in_group_p(op_data->op_attr.ia_gid))
rec->sa_suppgid =
from_kgid(&init_user_ns, op_data->op_attr.ia_gid);
else
rec->sa_suppgid = op_data->op_suppgids[0];
rec->sa_bias = op_data->op_bias;
}
/* WARNING: the file systems must take care not to tinker with
* attributes they don't manage (such as blocks).
*/
void obdo_from_inode(struct obdo *dst, struct inode *src, u32 valid)
{
u32 newvalid = 0;
if (valid & (OBD_MD_FLCTIME | OBD_MD_FLMTIME))
CDEBUG(D_INODE, "valid %x, new time %lu/%lu\n",
valid, LTIME_S(src->i_mtime),
LTIME_S(src->i_ctime));
if (valid & OBD_MD_FLATIME) {
dst->o_atime = LTIME_S(src->i_atime);
newvalid |= OBD_MD_FLATIME;
}
if (valid & OBD_MD_FLMTIME) {
dst->o_mtime = LTIME_S(src->i_mtime);
newvalid |= OBD_MD_FLMTIME;
}
if (valid & OBD_MD_FLCTIME) {
dst->o_ctime = LTIME_S(src->i_ctime);
newvalid |= OBD_MD_FLCTIME;
}
if (valid & OBD_MD_FLSIZE) {
dst->o_size = i_size_read(src);
newvalid |= OBD_MD_FLSIZE;
}
if (valid & OBD_MD_FLBLOCKS) { /* allocation of space (x512 bytes) */
dst->o_blocks = src->i_blocks;
newvalid |= OBD_MD_FLBLOCKS;
}
if (valid & OBD_MD_FLBLKSZ) { /* optimal block size */
dst->o_blksize = 1 << src->i_blkbits;
newvalid |= OBD_MD_FLBLKSZ;
}
if (valid & OBD_MD_FLTYPE) {
dst->o_mode = (dst->o_mode & S_IALLUGO) |
(src->i_mode & S_IFMT);
newvalid |= OBD_MD_FLTYPE;
}
if (valid & OBD_MD_FLMODE) {
dst->o_mode = (dst->o_mode & S_IFMT) |
(src->i_mode & S_IALLUGO);
newvalid |= OBD_MD_FLMODE;
}
if (valid & OBD_MD_FLUID) {
dst->o_uid = from_kuid(&init_user_ns, src->i_uid);
newvalid |= OBD_MD_FLUID;
}
if (valid & OBD_MD_FLGID) {
dst->o_gid = from_kgid(&init_user_ns, src->i_gid);
newvalid |= OBD_MD_FLGID;
}
if (valid & OBD_MD_FLFLAGS) {
dst->o_flags = src->i_flags;
newvalid |= OBD_MD_FLFLAGS;
}
dst->o_valid |= newvalid;
}
int mdc_setattr(struct obd_export *exp, struct md_op_data *op_data,
void *ea, size_t ealen, struct ptlrpc_request **request)
{
struct list_head cancels = LIST_HEAD_INIT(cancels);
struct ptlrpc_request *req;
struct mdc_rpc_lock *rpc_lock;
struct obd_device *obd = exp->exp_obd;
int count = 0, rc;
__u64 bits;
ENTRY;
LASSERT(op_data != NULL);
bits = MDS_INODELOCK_UPDATE;
if (op_data->op_attr.ia_valid & (ATTR_MODE|ATTR_UID|ATTR_GID))
bits |= MDS_INODELOCK_LOOKUP;
if ((op_data->op_flags & MF_MDC_CANCEL_FID1) &&
(fid_is_sane(&op_data->op_fid1)))
count = mdc_resource_get_unused(exp, &op_data->op_fid1,
&cancels, LCK_EX, bits);
req = ptlrpc_request_alloc(class_exp2cliimp(exp),
&RQF_MDS_REINT_SETATTR);
if (req == NULL) {
ldlm_lock_list_put(&cancels, l_bl_ast, count);
RETURN(-ENOMEM);
}
mdc_set_capa_size(req, &RMF_CAPA1, op_data->op_capa1);
req_capsule_set_size(&req->rq_pill, &RMF_MDT_EPOCH, RCL_CLIENT, 0);
req_capsule_set_size(&req->rq_pill, &RMF_EADATA, RCL_CLIENT, ealen);
req_capsule_set_size(&req->rq_pill, &RMF_LOGCOOKIES, RCL_CLIENT, 0);
rc = mdc_prep_elc_req(exp, req, MDS_REINT, &cancels, count);
if (rc) {
ptlrpc_request_free(req);
RETURN(rc);
}
rpc_lock = obd->u.cli.cl_rpc_lock;
if (op_data->op_attr.ia_valid & (ATTR_MTIME | ATTR_CTIME))
CDEBUG(D_INODE, "setting mtime "CFS_TIME_T
", ctime "CFS_TIME_T"\n",
LTIME_S(op_data->op_attr.ia_mtime),
LTIME_S(op_data->op_attr.ia_ctime));
mdc_setattr_pack(req, op_data, ea, ealen);
ptlrpc_request_set_replen(req);
rc = mdc_reint(req, rpc_lock, LUSTRE_IMP_FULL);
if (rc == -ERESTARTSYS)
rc = 0;
*request = req;
RETURN(rc);
}
void obdo_refresh_inode(struct inode *dst, struct obdo *src, obd_flag valid)
{
valid &= src->o_valid;
if (valid & (OBD_MD_FLCTIME | OBD_MD_FLMTIME))
CDEBUG(D_INODE,
"valid %#llx, cur time %lu/%lu, new %llu/%llu\n",
src->o_valid, LTIME_S(dst->i_mtime),
LTIME_S(dst->i_ctime), src->o_mtime, src->o_ctime);
if (valid & OBD_MD_FLATIME && src->o_atime > LTIME_S(dst->i_atime))
LTIME_S(dst->i_atime) = src->o_atime;
if (valid & OBD_MD_FLMTIME && src->o_mtime > LTIME_S(dst->i_mtime))
LTIME_S(dst->i_mtime) = src->o_mtime;
if (valid & OBD_MD_FLCTIME && src->o_ctime > LTIME_S(dst->i_ctime))
LTIME_S(dst->i_ctime) = src->o_ctime;
if (valid & OBD_MD_FLSIZE)
i_size_write(dst, src->o_size);
/* optimum IO size */
if (valid & OBD_MD_FLBLKSZ && src->o_blksize > (1 << dst->i_blkbits))
dst->i_blkbits = ffs(src->o_blksize) - 1;
if (dst->i_blkbits < PAGE_CACHE_SHIFT)
dst->i_blkbits = PAGE_CACHE_SHIFT;
/* allocation of space */
if (valid & OBD_MD_FLBLOCKS && src->o_blocks > dst->i_blocks)
/*
* XXX shouldn't overflow be checked here like in
* obdo_to_inode().
*/
dst->i_blocks = src->o_blocks;
}
void obdo_refresh_inode(struct inode *dst,
struct obdo *src,
obd_flag valid)
{
struct intnl_stat *st = llu_i2stat(dst);
valid &= src->o_valid;
if (valid & (OBD_MD_FLCTIME | OBD_MD_FLMTIME))
CDEBUG(D_INODE,"valid "LPX64", cur time "CFS_TIME_T"/"CFS_TIME_T
", new %lu/%lu\n",
src->o_valid, LTIME_S(st->st_mtime),
LTIME_S(st->st_ctime),
(long)src->o_mtime, (long)src->o_ctime);
if (valid & OBD_MD_FLATIME && src->o_atime > LTIME_S(st->st_atime))
LTIME_S(st->st_atime) = src->o_atime;
if (valid & OBD_MD_FLMTIME && src->o_mtime > LTIME_S(st->st_mtime))
LTIME_S(st->st_mtime) = src->o_mtime;
if (valid & OBD_MD_FLCTIME && src->o_ctime > LTIME_S(st->st_ctime))
LTIME_S(st->st_ctime) = src->o_ctime;
if (valid & OBD_MD_FLSIZE && src->o_size > st->st_size)
st->st_size = src->o_size;
/* optimum IO size */
if (valid & OBD_MD_FLBLKSZ)
st->st_blksize = src->o_blksize;
/* allocation of space */
if (valid & OBD_MD_FLBLOCKS && src->o_blocks > st->st_blocks)
st->st_blocks = src->o_blocks;
}
int mdc_setattr(struct obd_export *exp, struct md_op_data *op_data,
void *ea, size_t ealen, struct ptlrpc_request **request)
{
LIST_HEAD(cancels);
struct ptlrpc_request *req;
int count = 0, rc;
__u64 bits;
bits = MDS_INODELOCK_UPDATE;
if (op_data->op_attr.ia_valid & (ATTR_MODE | ATTR_UID | ATTR_GID))
bits |= MDS_INODELOCK_LOOKUP;
if ((op_data->op_flags & MF_MDC_CANCEL_FID1) &&
(fid_is_sane(&op_data->op_fid1)))
count = mdc_resource_get_unused(exp, &op_data->op_fid1,
&cancels, LCK_EX, bits);
req = ptlrpc_request_alloc(class_exp2cliimp(exp),
&RQF_MDS_REINT_SETATTR);
if (!req) {
ldlm_lock_list_put(&cancels, l_bl_ast, count);
return -ENOMEM;
}
req_capsule_set_size(&req->rq_pill, &RMF_MDT_EPOCH, RCL_CLIENT, 0);
req_capsule_set_size(&req->rq_pill, &RMF_EADATA, RCL_CLIENT, ealen);
req_capsule_set_size(&req->rq_pill, &RMF_LOGCOOKIES, RCL_CLIENT, 0);
rc = mdc_prep_elc_req(exp, req, MDS_REINT, &cancels, count);
if (rc) {
ptlrpc_request_free(req);
return rc;
}
if (op_data->op_attr.ia_valid & (ATTR_MTIME | ATTR_CTIME))
CDEBUG(D_INODE, "setting mtime %ld, ctime %ld\n",
LTIME_S(op_data->op_attr.ia_mtime),
LTIME_S(op_data->op_attr.ia_ctime));
mdc_setattr_pack(req, op_data, ea, ealen);
ptlrpc_request_set_replen(req);
rc = mdc_reint(req, LUSTRE_IMP_FULL);
if (rc == -ERESTARTSYS)
rc = 0;
*request = req;
return rc;
}
void iattr_from_obdo(struct iattr *attr, struct obdo *oa, obd_flag valid)
{
valid &= oa->o_valid;
if (valid & (OBD_MD_FLCTIME | OBD_MD_FLMTIME))
CDEBUG(D_INODE, "valid "LPX64", new time "LPU64"/"LPU64"\n",
oa->o_valid, oa->o_mtime, oa->o_ctime);
attr->ia_valid = 0;
if (valid & OBD_MD_FLATIME) {
LTIME_S(attr->ia_atime) = oa->o_atime;
attr->ia_valid |= ATTR_ATIME;
}
if (valid & OBD_MD_FLMTIME) {
LTIME_S(attr->ia_mtime) = oa->o_mtime;
attr->ia_valid |= ATTR_MTIME;
}
if (valid & OBD_MD_FLCTIME) {
LTIME_S(attr->ia_ctime) = oa->o_ctime;
attr->ia_valid |= ATTR_CTIME;
}
if (valid & OBD_MD_FLSIZE) {
attr->ia_size = oa->o_size;
attr->ia_valid |= ATTR_SIZE;
}
#if 0 /* you shouldn't be able to change a file's type with setattr */
if (valid & OBD_MD_FLTYPE) {
attr->ia_mode = (attr->ia_mode & ~S_IFMT)|(oa->o_mode & S_IFMT);
attr->ia_valid |= ATTR_MODE;
}
#endif
if (valid & OBD_MD_FLMODE) {
attr->ia_mode = (attr->ia_mode & S_IFMT)|(oa->o_mode & ~S_IFMT);
attr->ia_valid |= ATTR_MODE;
if (!in_group_p(make_kgid(&init_user_ns, oa->o_gid)) &&
!cfs_capable(CFS_CAP_FSETID))
attr->ia_mode &= ~S_ISGID;
}
if (valid & OBD_MD_FLUID) {
attr->ia_uid = make_kuid(&init_user_ns, oa->o_uid);
attr->ia_valid |= ATTR_UID;
}
if (valid & OBD_MD_FLGID) {
attr->ia_gid = make_kgid(&init_user_ns, oa->o_gid);
attr->ia_valid |= ATTR_GID;
}
}
static int vvp_object_glimpse(const struct lu_env *env,
const struct cl_object *obj, struct ost_lvb *lvb)
{
struct inode *inode = vvp_object_inode(obj);
lvb->lvb_mtime = LTIME_S(inode->i_mtime);
lvb->lvb_atime = LTIME_S(inode->i_atime);
lvb->lvb_ctime = LTIME_S(inode->i_ctime);
/*
* LU-417: Add dirty pages block count lest i_blocks reports 0, some
* "cp" or "tar" on remote node may think it's a completely sparse file
* and skip it.
*/
if (lvb->lvb_size > 0 && lvb->lvb_blocks == 0)
lvb->lvb_blocks = dirty_cnt(inode);
return 0;
}
static int vvp_io_fault_iter_init(const struct lu_env *env,
const struct cl_io_slice *ios)
{
struct vvp_io *vio = cl2vvp_io(env, ios);
struct inode *inode = ccc_object_inode(ios->cis_obj);
LASSERT(inode ==
cl2ccc_io(env, ios)->cui_fd->fd_file->f_dentry->d_inode);
vio->u.fault.ft_mtime = LTIME_S(inode->i_mtime);
return 0;
}
static int vvp_attr_get(const struct lu_env *env, struct cl_object *obj,
struct cl_attr *attr)
{
struct inode *inode = vvp_object_inode(obj);
/*
* lov overwrites most of these fields in
* lov_attr_get()->...lov_merge_lvb_kms(), except when inode
* attributes are newer.
*/
attr->cat_size = i_size_read(inode);
attr->cat_mtime = LTIME_S(inode->i_mtime);
attr->cat_atime = LTIME_S(inode->i_atime);
attr->cat_ctime = LTIME_S(inode->i_ctime);
attr->cat_blocks = inode->i_blocks;
attr->cat_uid = from_kuid(&init_user_ns, inode->i_uid);
attr->cat_gid = from_kgid(&init_user_ns, inode->i_gid);
/* KMS is not known by this layer */
return 0; /* layers below have to fill in the rest */
}
static int vvp_attr_set(const struct lu_env *env, struct cl_object *obj,
const struct cl_attr *attr, unsigned valid)
{
struct inode *inode = ccc_object_inode(obj);
if (valid & CAT_UID)
inode->i_uid = attr->cat_uid;
if (valid & CAT_GID)
inode->i_gid = attr->cat_gid;
if (valid & CAT_ATIME)
LTIME_S(inode->i_atime) = attr->cat_atime;
if (valid & CAT_MTIME)
LTIME_S(inode->i_mtime) = attr->cat_mtime;
if (valid & CAT_CTIME)
LTIME_S(inode->i_ctime) = attr->cat_ctime;
if (0 && valid & CAT_SIZE)
cl_isize_write_nolock(inode, attr->cat_size);
/* not currently necessary */
if (0 && valid & (CAT_UID|CAT_GID|CAT_SIZE))
mark_inode_dirty(inode);
return 0;
}
static int vvp_attr_update(const struct lu_env *env, struct cl_object *obj,
const struct cl_attr *attr, unsigned valid)
{
struct inode *inode = vvp_object_inode(obj);
if (valid & CAT_UID)
inode->i_uid = make_kuid(&init_user_ns, attr->cat_uid);
if (valid & CAT_GID)
inode->i_gid = make_kgid(&init_user_ns, attr->cat_gid);
if (valid & CAT_ATIME)
LTIME_S(inode->i_atime) = attr->cat_atime;
if (valid & CAT_MTIME)
LTIME_S(inode->i_mtime) = attr->cat_mtime;
if (valid & CAT_CTIME)
LTIME_S(inode->i_ctime) = attr->cat_ctime;
if (0 && valid & CAT_SIZE)
i_size_write(inode, attr->cat_size);
/* not currently necessary */
if (0 && valid & (CAT_UID|CAT_GID|CAT_SIZE))
mark_inode_dirty(inode);
return 0;
}
static int osc_io_read_start(const struct lu_env *env,
const struct cl_io_slice *slice)
{
struct cl_object *obj = slice->cis_obj;
struct cl_attr *attr = &osc_env_info(env)->oti_attr;
int rc = 0;
if (!slice->cis_io->ci_noatime) {
cl_object_attr_lock(obj);
attr->cat_atime = LTIME_S(CURRENT_TIME);
rc = cl_object_attr_set(env, obj, attr, CAT_ATIME);
cl_object_attr_unlock(obj);
}
return rc;
}
static int osc_io_write_start(const struct lu_env *env,
const struct cl_io_slice *slice)
{
struct cl_object *obj = slice->cis_obj;
struct cl_attr *attr = &osc_env_info(env)->oti_attr;
int rc = 0;
OBD_FAIL_TIMEOUT(OBD_FAIL_OSC_DELAY_SETTIME, 1);
cl_object_attr_lock(obj);
attr->cat_mtime = attr->cat_ctime = LTIME_S(CURRENT_TIME);
rc = cl_object_attr_set(env, obj, attr, CAT_MTIME | CAT_CTIME);
cl_object_attr_unlock(obj);
return rc;
}
static int osc_io_read_start(const struct lu_env *env,
const struct cl_io_slice *slice)
{
struct osc_io *oio = cl2osc_io(env, slice);
struct cl_object *obj = slice->cis_obj;
struct cl_attr *attr = &osc_env_info(env)->oti_attr;
int rc = 0;
ENTRY;
if (oio->oi_lockless == 0 && !slice->cis_io->ci_noatime) {
cl_object_attr_lock(obj);
attr->cat_atime = LTIME_S(CFS_CURRENT_TIME);
rc = cl_object_attr_set(env, obj, attr, CAT_ATIME);
cl_object_attr_unlock(obj);
}
RETURN(rc);
}
static int osc_io_write_start(const struct lu_env *env,
const struct cl_io_slice *slice)
{
struct osc_io *oio = cl2osc_io(env, slice);
struct cl_object *obj = slice->cis_obj;
struct cl_attr *attr = &osc_env_info(env)->oti_attr;
int result = 0;
ENTRY;
if (oio->oi_lockless == 0) {
OBD_FAIL_TIMEOUT(OBD_FAIL_OSC_DELAY_SETTIME, 1);
cl_object_attr_lock(obj);
result = cl_object_attr_get(env, obj, attr);
if (result == 0) {
attr->cat_mtime = attr->cat_ctime =
LTIME_S(CFS_CURRENT_TIME);
result = cl_object_attr_set(env, obj, attr,
CAT_MTIME | CAT_CTIME);
}
cl_object_attr_unlock(obj);
}
RETURN(result);
}
/**
* This is called when a page is accessed within file in a way that creates
* new page, if one were missing (i.e., if there were a hole at that place in
* the file, or accessed page is beyond the current file size).
*
* Expand stripe KMS if necessary.
*/
static void osc_page_touch_at(const struct lu_env *env,
struct cl_object *obj, pgoff_t idx, size_t to)
{
struct lov_oinfo *loi = cl2osc(obj)->oo_oinfo;
struct cl_attr *attr = &osc_env_info(env)->oti_attr;
int valid;
__u64 kms;
/* offset within stripe */
kms = cl_offset(obj, idx) + to;
cl_object_attr_lock(obj);
/*
* XXX old code used
*
* ll_inode_size_lock(inode, 0); lov_stripe_lock(lsm);
*
* here
*/
CDEBUG(D_INODE, "stripe KMS %sincreasing "LPU64"->"LPU64" "LPU64"\n",
kms > loi->loi_kms ? "" : "not ", loi->loi_kms, kms,
loi->loi_lvb.lvb_size);
attr->cat_mtime = attr->cat_ctime = LTIME_S(CFS_CURRENT_TIME);
valid = CAT_MTIME | CAT_CTIME;
if (kms > loi->loi_kms) {
attr->cat_kms = kms;
valid |= CAT_KMS;
}
if (kms > loi->loi_lvb.lvb_size) {
attr->cat_size = kms;
valid |= CAT_SIZE;
}
cl_object_attr_update(env, obj, attr, valid);
cl_object_attr_unlock(obj);
}
void obdo_to_inode(struct inode *dst, struct obdo *src, obd_flag valid)
{
valid &= src->o_valid;
LASSERTF(!(valid & (OBD_MD_FLTYPE | OBD_MD_FLGENER | OBD_MD_FLFID |
OBD_MD_FLID | OBD_MD_FLGROUP)),
"object "DOSTID", valid %x\n", POSTID(&src->o_oi), valid);
if (valid & (OBD_MD_FLCTIME | OBD_MD_FLMTIME))
CDEBUG(D_INODE,
"valid %#llx, cur time %lu/%lu, new %llu/%llu\n",
src->o_valid, LTIME_S(dst->i_mtime),
LTIME_S(dst->i_ctime), src->o_mtime, src->o_ctime);
if (valid & OBD_MD_FLATIME)
LTIME_S(dst->i_atime) = src->o_atime;
if (valid & OBD_MD_FLMTIME)
LTIME_S(dst->i_mtime) = src->o_mtime;
if (valid & OBD_MD_FLCTIME && src->o_ctime > LTIME_S(dst->i_ctime))
LTIME_S(dst->i_ctime) = src->o_ctime;
if (valid & OBD_MD_FLSIZE)
i_size_write(dst, src->o_size);
if (valid & OBD_MD_FLBLOCKS) { /* allocation of space */
dst->i_blocks = src->o_blocks;
if (dst->i_blocks < src->o_blocks) /* overflow */
dst->i_blocks = -1;
}
if (valid & OBD_MD_FLBLKSZ)
dst->i_blkbits = ffs(src->o_blksize)-1;
if (valid & OBD_MD_FLMODE)
dst->i_mode = (dst->i_mode & S_IFMT) | (src->o_mode & ~S_IFMT);
if (valid & OBD_MD_FLUID)
dst->i_uid = make_kuid(&init_user_ns, src->o_uid);
if (valid & OBD_MD_FLGID)
dst->i_gid = make_kgid(&init_user_ns, src->o_gid);
if (valid & OBD_MD_FLFLAGS)
dst->i_flags = src->o_flags;
}
static int vvp_io_fault_start(const struct lu_env *env,
const struct cl_io_slice *ios)
{
struct vvp_io *vio = cl2vvp_io(env, ios);
struct cl_io *io = ios->cis_io;
struct cl_object *obj = io->ci_obj;
struct inode *inode = ccc_object_inode(obj);
struct cl_fault_io *fio = &io->u.ci_fault;
struct vvp_fault_io *cfio = &vio->u.fault;
loff_t offset;
int result = 0;
struct page *vmpage = NULL;
struct cl_page *page;
loff_t size;
pgoff_t last; /* last page in a file data region */
if (fio->ft_executable &&
LTIME_S(inode->i_mtime) != vio->u.fault.ft_mtime)
CWARN("binary "DFID
" changed while waiting for the page fault lock\n",
PFID(lu_object_fid(&obj->co_lu)));
/* offset of the last byte on the page */
offset = cl_offset(obj, fio->ft_index + 1) - 1;
LASSERT(cl_index(obj, offset) == fio->ft_index);
result = ccc_prep_size(env, obj, io, 0, offset + 1, NULL);
if (result != 0)
return result;
/* must return locked page */
if (fio->ft_mkwrite) {
LASSERT(cfio->ft_vmpage != NULL);
lock_page(cfio->ft_vmpage);
} else {
result = vvp_io_kernel_fault(cfio);
if (result != 0)
return result;
}
vmpage = cfio->ft_vmpage;
LASSERT(PageLocked(vmpage));
if (OBD_FAIL_CHECK(OBD_FAIL_LLITE_FAULT_TRUNC_RACE))
ll_invalidate_page(vmpage);
size = i_size_read(inode);
/* Though we have already held a cl_lock upon this page, but
* it still can be truncated locally. */
if (unlikely((vmpage->mapping != inode->i_mapping) ||
(page_offset(vmpage) > size))) {
CDEBUG(D_PAGE, "llite: fault and truncate race happened!\n");
/* return +1 to stop cl_io_loop() and ll_fault() will catch
* and retry. */
GOTO(out, result = +1);
}
if (fio->ft_mkwrite ) {
pgoff_t last_index;
/*
* Capture the size while holding the lli_trunc_sem from above
* we want to make sure that we complete the mkwrite action
* while holding this lock. We need to make sure that we are
* not past the end of the file.
*/
last_index = cl_index(obj, size - 1);
if (last_index < fio->ft_index) {
CDEBUG(D_PAGE,
"llite: mkwrite and truncate race happened: "
"%p: 0x%lx 0x%lx\n",
vmpage->mapping,fio->ft_index,last_index);
/*
* We need to return if we are
* passed the end of the file. This will propagate
* up the call stack to ll_page_mkwrite where
* we will return VM_FAULT_NOPAGE. Any non-negative
* value returned here will be silently
* converted to 0. If the vmpage->mapping is null
* the error code would be converted back to ENODATA
* in ll_page_mkwrite0. Thus we return -ENODATA
* to handle both cases
*/
GOTO(out, result = -ENODATA);
}
}
page = cl_page_find(env, obj, fio->ft_index, vmpage, CPT_CACHEABLE);
if (IS_ERR(page))
GOTO(out, result = PTR_ERR(page));
/* if page is going to be written, we should add this page into cache
* earlier. */
if (fio->ft_mkwrite) {
wait_on_page_writeback(vmpage);
if (set_page_dirty(vmpage)) {
struct ccc_page *cp;
/* vvp_page_assume() calls wait_on_page_writeback(). */
cl_page_assume(env, io, page);
cp = cl2ccc_page(cl_page_at(page, &vvp_device_type));
vvp_write_pending(cl2ccc(obj), cp);
/* Do not set Dirty bit here so that in case IO is
* started before the page is really made dirty, we
* still have chance to detect it. */
result = cl_page_cache_add(env, io, page, CRT_WRITE);
LASSERT(cl_page_is_owned(page, io));
vmpage = NULL;
if (result < 0) {
cl_page_unmap(env, io, page);
cl_page_discard(env, io, page);
cl_page_disown(env, io, page);
cl_page_put(env, page);
/* we're in big trouble, what can we do now? */
if (result == -EDQUOT)
result = -ENOSPC;
GOTO(out, result);
} else
cl_page_disown(env, io, page);
}
}
last = cl_index(obj, size - 1);
/*
* The ft_index is only used in the case of
* a mkwrite action. We need to check
* our assertions are correct, since
* we should have caught this above
*/
LASSERT(!fio->ft_mkwrite || fio->ft_index <= last);
if (fio->ft_index == last)
/*
* Last page is mapped partially.
*/
fio->ft_nob = size - cl_offset(obj, fio->ft_index);
else
fio->ft_nob = cl_page_size(obj);
lu_ref_add(&page->cp_reference, "fault", io);
fio->ft_page = page;
EXIT;
out:
/* return unlocked vmpage to avoid deadlocking */
if (vmpage != NULL)
unlock_page(vmpage);
cfio->fault.ft_flags &= ~VM_FAULT_LOCKED;
return result;
}
int lmv_revalidate_slaves(struct obd_export *exp,
const struct lmv_stripe_md *lsm,
ldlm_blocking_callback cb_blocking,
int extra_lock_flags)
{
struct obd_device *obd = exp->exp_obd;
struct lmv_obd *lmv = &obd->u.lmv;
struct ptlrpc_request *req = NULL;
struct mdt_body *body;
struct md_op_data *op_data;
int i;
int rc = 0;
ENTRY;
/**
* revalidate slaves has some problems, temporarily return,
* we may not need that
*/
OBD_ALLOC_PTR(op_data);
if (op_data == NULL)
RETURN(-ENOMEM);
/**
* Loop over the stripe information, check validity and update them
* from MDS if needed.
*/
for (i = 0; i < lsm->lsm_md_stripe_count; i++) {
struct lu_fid fid;
struct lookup_intent it = { .it_op = IT_GETATTR };
struct lustre_handle *lockh = NULL;
struct lmv_tgt_desc *tgt = NULL;
struct inode *inode;
fid = lsm->lsm_md_oinfo[i].lmo_fid;
inode = lsm->lsm_md_oinfo[i].lmo_root;
/*
* Prepare op_data for revalidating. Note that @fid2 shluld be
* defined otherwise it will go to server and take new lock
* which is not needed here.
*/
memset(op_data, 0, sizeof(*op_data));
op_data->op_fid1 = fid;
op_data->op_fid2 = fid;
tgt = lmv_locate_mds(lmv, op_data, &fid);
if (IS_ERR(tgt))
GOTO(cleanup, rc = PTR_ERR(tgt));
CDEBUG(D_INODE, "Revalidate slave "DFID" -> mds #%u\n",
PFID(&fid), tgt->ltd_idx);
if (req != NULL) {
ptlrpc_req_finished(req);
req = NULL;
}
rc = md_intent_lock(tgt->ltd_exp, op_data, &it, &req,
cb_blocking, extra_lock_flags);
if (rc < 0)
GOTO(cleanup, rc);
lockh = (struct lustre_handle *)&it.d.lustre.it_lock_handle;
if (rc > 0 && req == NULL) {
/* slave inode is still valid */
CDEBUG(D_INODE, "slave "DFID" is still valid.\n",
PFID(&fid));
rc = 0;
} else {
/* refresh slave from server */
body = req_capsule_server_get(&req->rq_pill,
&RMF_MDT_BODY);
if (body == NULL) {
if (it.d.lustre.it_lock_mode && lockh) {
ldlm_lock_decref(lockh,
it.d.lustre.it_lock_mode);
it.d.lustre.it_lock_mode = 0;
}
GOTO(cleanup, rc = -ENOENT);
}
i_size_write(inode, body->mbo_size);
inode->i_blocks = body->mbo_blocks;
set_nlink(inode, body->mbo_nlink);
LTIME_S(inode->i_atime) = body->mbo_atime;
LTIME_S(inode->i_ctime) = body->mbo_ctime;
LTIME_S(inode->i_mtime) = body->mbo_mtime;
}
md_set_lock_data(tgt->ltd_exp, &lockh->cookie, inode, NULL);
if (it.d.lustre.it_lock_mode != 0 && lockh != NULL) {
ldlm_lock_decref(lockh, it.d.lustre.it_lock_mode);
it.d.lustre.it_lock_mode = 0;
}
}
cleanup:
if (req != NULL)
ptlrpc_req_finished(req);
OBD_FREE_PTR(op_data);
RETURN(rc);
}
/* If mdc_setattr is called with an 'iattr', then it is a normal RPC that
* should take the normal semaphore and go to the normal portal.
*
* If it is called with iattr->ia_valid & ATTR_FROM_OPEN, then it is a
* magic open-path setattr that should take the setattr semaphore and
* go to the setattr portal. */
int mdc_setattr(struct obd_export *exp, struct md_op_data *op_data,
void *ea, int ealen, void *ea2, int ea2len,
struct ptlrpc_request **request, struct md_open_data **mod)
{
CFS_LIST_HEAD(cancels);
struct ptlrpc_request *req;
struct mdc_rpc_lock *rpc_lock;
struct obd_device *obd = exp->exp_obd;
int count = 0, rc;
__u64 bits;
ENTRY;
LASSERT(op_data != NULL);
bits = MDS_INODELOCK_UPDATE;
if (op_data->op_attr.ia_valid & (ATTR_MODE|ATTR_UID|ATTR_GID))
bits |= MDS_INODELOCK_LOOKUP;
if ((op_data->op_flags & MF_MDC_CANCEL_FID1) &&
(fid_is_sane(&op_data->op_fid1)))
count = mdc_resource_get_unused(exp, &op_data->op_fid1,
&cancels, LCK_EX, bits);
req = ptlrpc_request_alloc(class_exp2cliimp(exp),
&RQF_MDS_REINT_SETATTR);
if (req == NULL) {
ldlm_lock_list_put(&cancels, l_bl_ast, count);
RETURN(-ENOMEM);
}
mdc_set_capa_size(req, &RMF_CAPA1, op_data->op_capa1);
if ((op_data->op_flags & (MF_SOM_CHANGE | MF_EPOCH_OPEN)) == 0)
req_capsule_set_size(&req->rq_pill, &RMF_MDT_EPOCH, RCL_CLIENT,
0);
req_capsule_set_size(&req->rq_pill, &RMF_EADATA, RCL_CLIENT, ealen);
req_capsule_set_size(&req->rq_pill, &RMF_LOGCOOKIES, RCL_CLIENT,
ea2len);
rc = mdc_prep_elc_req(exp, req, &cancels, count);
if (rc) {
ptlrpc_request_free(req);
RETURN(rc);
}
if (op_data->op_attr.ia_valid & ATTR_FROM_OPEN) {
req->rq_request_portal = MDS_SETATTR_PORTAL;
ptlrpc_at_set_req_timeout(req);
rpc_lock = obd->u.cli.cl_setattr_lock;
} else {
rpc_lock = obd->u.cli.cl_rpc_lock;
}
if (op_data->op_attr.ia_valid & (ATTR_MTIME | ATTR_CTIME))
CDEBUG(D_INODE, "setting mtime "CFS_TIME_T
", ctime "CFS_TIME_T"\n",
LTIME_S(op_data->op_attr.ia_mtime),
LTIME_S(op_data->op_attr.ia_ctime));
mdc_setattr_pack(req, op_data, ea, ealen, ea2, ea2len);
ptlrpc_request_set_replen(req);
if (mod && (op_data->op_flags & MF_EPOCH_OPEN) &&
req->rq_import->imp_replayable)
{
LASSERT(*mod == NULL);
*mod = obd_mod_alloc();
if (*mod == NULL) {
DEBUG_REQ(D_ERROR, req, "Can't allocate "
"md_open_data");
} else {
req->rq_replay = 1;
req->rq_cb_data = *mod;
(*mod)->mod_open_req = req;
req->rq_commit_cb = mdc_commit_open;
/**
* Take an extra reference on \var mod, it protects \var
* mod from being freed on eviction (commit callback is
* called despite rq_replay flag).
* Will be put on mdc_done_writing().
*/
obd_mod_get(*mod);
}
}
rc = mdc_reint(req, rpc_lock, LUSTRE_IMP_FULL);
/* Save the obtained info in the original RPC for the replay case. */
if (rc == 0 && (op_data->op_flags & MF_EPOCH_OPEN)) {
struct mdt_ioepoch *epoch;
struct mdt_body *body;
epoch = req_capsule_client_get(&req->rq_pill, &RMF_MDT_EPOCH);
body = req_capsule_server_get(&req->rq_pill, &RMF_MDT_BODY);
LASSERT(epoch != NULL);
LASSERT(body != NULL);
epoch->handle = body->handle;
epoch->ioepoch = body->ioepoch;
req->rq_replay_cb = mdc_replay_open;
/** bug 3633, open may be committed and estale answer is not error */
} else if (rc == -ESTALE && (op_data->op_flags & MF_SOM_CHANGE)) {
rc = 0;
} else if (rc == -ERESTARTSYS) {
rc = 0;
}
*request = req;
if (rc && req->rq_commit_cb) {
/* Put an extra reference on \var mod on error case. */
obd_mod_put(*mod);
req->rq_commit_cb(req);
}
RETURN(rc);
}
int lmv_revalidate_slaves(struct obd_export *exp, struct mdt_body *mbody,
struct lmv_stripe_md *lsm,
ldlm_blocking_callback cb_blocking,
int extra_lock_flags)
{
struct obd_device *obd = exp->exp_obd;
struct lmv_obd *lmv = &obd->u.lmv;
struct mdt_body *body;
struct md_op_data *op_data;
unsigned long size = 0;
unsigned long nlink = 0;
obd_time atime = 0;
obd_time ctime = 0;
obd_time mtime = 0;
int i;
int rc = 0;
ENTRY;
/**
* revalidate slaves has some problems, temporarily return,
* we may not need that
*/
if (lsm->lsm_md_stripe_count <= 1)
RETURN(0);
OBD_ALLOC_PTR(op_data);
if (op_data == NULL)
RETURN(-ENOMEM);
/**
* Loop over the stripe information, check validity and update them
* from MDS if needed.
*/
for (i = 0; i < lsm->lsm_md_stripe_count; i++) {
struct lu_fid fid;
struct lookup_intent it = { .it_op = IT_GETATTR };
struct ptlrpc_request *req = NULL;
struct lustre_handle *lockh = NULL;
struct lmv_tgt_desc *tgt = NULL;
struct inode *inode;
fid = lsm->lsm_md_oinfo[i].lmo_fid;
inode = lsm->lsm_md_oinfo[i].lmo_root;
if (i == 0) {
if (mbody != NULL) {
body = mbody;
goto update;
} else {
goto release_lock;
}
}
/*
* Prepare op_data for revalidating. Note that @fid2 shluld be
* defined otherwise it will go to server and take new lock
* which is not needed here.
*/
memset(op_data, 0, sizeof(*op_data));
op_data->op_fid1 = fid;
op_data->op_fid2 = fid;
tgt = lmv_locate_mds(lmv, op_data, &fid);
if (IS_ERR(tgt))
GOTO(cleanup, rc = PTR_ERR(tgt));
CDEBUG(D_INODE, "Revalidate slave "DFID" -> mds #%d\n",
PFID(&fid), tgt->ltd_idx);
rc = md_intent_lock(tgt->ltd_exp, op_data, NULL, 0, &it, 0,
&req, cb_blocking, extra_lock_flags);
if (rc < 0)
GOTO(cleanup, rc);
lockh = (struct lustre_handle *)&it.d.lustre.it_lock_handle;
if (rc > 0 && req == NULL) {
/* slave inode is still valid */
CDEBUG(D_INODE, "slave "DFID" is still valid.\n",
PFID(&fid));
rc = 0;
} else {
/* refresh slave from server */
body = req_capsule_server_get(&req->rq_pill,
&RMF_MDT_BODY);
LASSERT(body != NULL);
update:
if (unlikely(body->nlink < 2)) {
CERROR("%s: nlink %d < 2 corrupt stripe %d "DFID
":" DFID"\n", obd->obd_name, body->nlink,
i, PFID(&lsm->lsm_md_oinfo[i].lmo_fid),
PFID(&lsm->lsm_md_oinfo[0].lmo_fid));
if (req != NULL)
ptlrpc_req_finished(req);
if (it.d.lustre.it_lock_mode && lockh) {
ldlm_lock_decref(lockh,
it.d.lustre.it_lock_mode);
it.d.lustre.it_lock_mode = 0;
}
GOTO(cleanup, rc = -EIO);
}
if (i != 0)
md_set_lock_data(tgt->ltd_exp, &lockh->cookie,
inode, NULL);
i_size_write(inode, body->size);
set_nlink(inode, body->nlink);
LTIME_S(inode->i_atime) = body->atime;
LTIME_S(inode->i_ctime) = body->ctime;
LTIME_S(inode->i_mtime) = body->mtime;
if (req != NULL)
ptlrpc_req_finished(req);
}
release_lock:
size += i_size_read(inode);
if (i != 0)
nlink += inode->i_nlink - 2;
else
nlink += inode->i_nlink;
atime = LTIME_S(inode->i_atime) > atime ?
LTIME_S(inode->i_atime) : atime;
ctime = LTIME_S(inode->i_ctime) > ctime ?
LTIME_S(inode->i_ctime) : ctime;
mtime = LTIME_S(inode->i_mtime) > mtime ?
LTIME_S(inode->i_mtime) : mtime;
if (it.d.lustre.it_lock_mode != 0 && lockh != NULL) {
ldlm_lock_decref(lockh, it.d.lustre.it_lock_mode);
it.d.lustre.it_lock_mode = 0;
}
CDEBUG(D_INODE, "i %d "DFID" size %llu, nlink %u, atime "
"%lu, mtime %lu, ctime %lu.\n", i, PFID(&fid),
i_size_read(inode), inode->i_nlink,
LTIME_S(inode->i_atime), LTIME_S(inode->i_mtime),
LTIME_S(inode->i_ctime));
}
/*
* update attr of master request.
*/
CDEBUG(D_INODE, "Return refreshed attrs: size = %lu nlink %lu atime "
LPU64 "ctime "LPU64" mtime "LPU64" for " DFID"\n", size, nlink,
atime, ctime, mtime, PFID(&lsm->lsm_md_oinfo[0].lmo_fid));
if (mbody != NULL) {
mbody->atime = atime;
mbody->ctime = ctime;
mbody->mtime = mtime;
}
cleanup:
OBD_FREE_PTR(op_data);
RETURN(rc);
}