static struct inode *ll_alloc_inode(struct super_block *sb)
{
struct ll_inode_info *lli;
ll_stats_ops_tally(ll_s2sbi(sb), LPROC_LL_ALLOC_INODE, 1);
OBD_SLAB_ALLOC_PTR_GFP(lli, ll_inode_cachep, CFS_ALLOC_IO);
if (lli == NULL)
return NULL;
inode_init_once(&lli->lli_vfs_inode);
return &lli->lli_vfs_inode;
}
int ll_setxattr(struct dentry *dentry, const char *name,
const void *value, size_t size, int flags)
{
struct inode *inode = d_inode(dentry);
LASSERT(inode);
LASSERT(name);
CDEBUG(D_VFSTRACE, "VFS Op:inode=%lu/%u(%p), xattr %s\n",
inode->i_ino, inode->i_generation, inode, name);
ll_stats_ops_tally(ll_i2sbi(inode), LPROC_LL_SETXATTR, 1);
if ((strncmp(name, XATTR_TRUSTED_PREFIX,
sizeof(XATTR_TRUSTED_PREFIX) - 1) == 0 &&
strcmp(name + sizeof(XATTR_TRUSTED_PREFIX) - 1, "lov") == 0) ||
(strncmp(name, XATTR_LUSTRE_PREFIX,
sizeof(XATTR_LUSTRE_PREFIX) - 1) == 0 &&
strcmp(name + sizeof(XATTR_LUSTRE_PREFIX) - 1, "lov") == 0)) {
struct lov_user_md *lump = (struct lov_user_md *)value;
int rc = 0;
if (size != 0 && size < sizeof(struct lov_user_md))
return -EINVAL;
/* Attributes that are saved via getxattr will always have
* the stripe_offset as 0. Instead, the MDS should be
* allowed to pick the starting OST index. b=17846
*/
if (lump && lump->lmm_stripe_offset == 0)
lump->lmm_stripe_offset = -1;
if (lump && S_ISREG(inode->i_mode)) {
int flags = FMODE_WRITE;
int lum_size = (lump->lmm_magic == LOV_USER_MAGIC_V1) ?
sizeof(*lump) : sizeof(struct lov_user_md_v3);
rc = ll_lov_setstripe_ea_info(inode, dentry, flags, lump,
lum_size);
/* b10667: rc always be 0 here for now */
rc = 0;
} else if (S_ISDIR(inode->i_mode)) {
rc = ll_dir_setstripe(inode, lump, 0);
}
return rc;
} else if (strcmp(name, XATTR_NAME_LMA) == 0 ||
strcmp(name, XATTR_NAME_LINK) == 0)
return 0;
return ll_setxattr_common(inode, name, value, size, flags,
OBD_MD_FLXATTR);
}
static struct inode *ll_alloc_inode(struct super_block *sb)
{
struct ll_inode_info *lli;
ll_stats_ops_tally(ll_s2sbi(sb), LPROC_LL_ALLOC_INODE, 1);
lli = kmem_cache_zalloc(ll_inode_cachep, GFP_NOFS);
if (!lli)
return NULL;
inode_init_once(&lli->lli_vfs_inode);
return &lli->lli_vfs_inode;
}
int ll_removexattr(struct dentry *dentry, const char *name)
{
struct inode *inode = dentry->d_inode;
LASSERT(inode);
LASSERT(name);
CDEBUG(D_VFSTRACE, "VFS Op:inode=%lu/%u(%p), xattr %s\n",
inode->i_ino, inode->i_generation, inode, name);
ll_stats_ops_tally(ll_i2sbi(inode), LPROC_LL_REMOVEXATTR, 1);
return ll_setxattr_common(inode, name, NULL, 0, 0,
OBD_MD_FLXATTRRM);
}
static int ll_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf)
{
int count = 0;
bool printed = false;
bool retry;
int result;
ll_stats_ops_tally(ll_i2sbi(file_inode(vma->vm_file)),
LPROC_LL_MKWRITE, 1);
file_update_time(vma->vm_file);
do {
retry = false;
result = ll_page_mkwrite0(vma, vmf->page, &retry);
if (!printed && ++count > 16) {
const struct dentry *de = file_dentry(vma->vm_file);
CWARN("app(%s): the page %lu of file "DFID" is under"
" heavy contention\n",
current->comm, vmf->pgoff,
PFID(ll_inode2fid(de->d_inode)));
printed = true;
}
} while (retry);
switch(result) {
case 0:
LASSERT(PageLocked(vmf->page));
result = VM_FAULT_LOCKED;
break;
case -ENODATA:
case -EFAULT:
result = VM_FAULT_NOPAGE;
break;
case -ENOMEM:
result = VM_FAULT_OOM;
break;
case -EAGAIN:
result = VM_FAULT_RETRY;
break;
default:
result = VM_FAULT_SIGBUS;
break;
}
return result;
}
int ll_file_mmap(struct file *file, struct vm_area_struct *vma)
{
struct inode *inode = file_inode(file);
int rc;
if (ll_file_nolock(file))
return -EOPNOTSUPP;
ll_stats_ops_tally(ll_i2sbi(inode), LPROC_LL_MAP, 1);
rc = generic_file_mmap(file, vma);
if (rc == 0) {
vma->vm_ops = &ll_file_vm_ops;
vma->vm_ops->open(vma);
/* update the inode's size and mtime */
rc = ll_glimpse_size(inode);
}
return rc;
}
static int ll_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
{
int count = 0;
bool printed = false;
int result;
sigset_t set;
/* Only SIGKILL and SIGTERM is allowed for fault/nopage/mkwrite
* so that it can be killed by admin but not cause segfault by
* other signals. */
set = cfs_block_sigsinv(sigmask(SIGKILL) | sigmask(SIGTERM));
ll_stats_ops_tally(ll_i2sbi(file_inode(vma->vm_file)),
LPROC_LL_FAULT, 1);
restart:
result = ll_fault0(vma, vmf);
if (!(result & (VM_FAULT_RETRY | VM_FAULT_ERROR | VM_FAULT_LOCKED))) {
struct page *vmpage = vmf->page;
/* check if this page has been truncated */
lock_page(vmpage);
if (unlikely(vmpage->mapping == NULL)) { /* unlucky */
unlock_page(vmpage);
put_page(vmpage);
vmf->page = NULL;
if (!printed && ++count > 16) {
CWARN("the page is under heavy contention,"
"maybe your app(%s) needs revising :-)\n",
current->comm);
printed = true;
}
goto restart;
}
result |= VM_FAULT_LOCKED;
}
cfs_restore_sigs(set);
return result;
}
ssize_t ll_listxattr(struct dentry *dentry, char *buffer, size_t size)
{
struct inode *inode = dentry->d_inode;
int rc = 0, rc2 = 0;
struct lov_mds_md *lmm = NULL;
struct ptlrpc_request *request = NULL;
int lmmsize;
LASSERT(inode);
CDEBUG(D_VFSTRACE, "VFS Op:inode=%lu/%u(%p)\n",
inode->i_ino, inode->i_generation, inode);
ll_stats_ops_tally(ll_i2sbi(inode), LPROC_LL_LISTXATTR, 1);
rc = ll_getxattr_common(inode, NULL, buffer, size, OBD_MD_FLXATTRLS);
if (rc < 0)
GOTO(out, rc);
if (buffer != NULL) {
struct ll_sb_info *sbi = ll_i2sbi(inode);
char *xattr_name = buffer;
int xlen, rem = rc;
while (rem > 0) {
xlen = strnlen(xattr_name, rem - 1) + 1;
rem -= xlen;
if (xattr_type_filter(sbi,
get_xattr_type(xattr_name)) == 0) {
/* skip OK xattr type
* leave it in buffer
*/
xattr_name += xlen;
continue;
}
/* move up remaining xattrs in buffer
* removing the xattr that is not OK
*/
memmove(xattr_name, xattr_name + xlen, rem);
rc -= xlen;
}
}
if (S_ISREG(inode->i_mode)) {
if (!ll_i2info(inode)->lli_has_smd)
rc2 = -1;
} else if (S_ISDIR(inode->i_mode)) {
rc2 = ll_dir_getstripe(inode, &lmm, &lmmsize, &request);
}
if (rc2 < 0) {
GOTO(out, rc2 = 0);
} else if (S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode)) {
const int prefix_len = sizeof(XATTR_LUSTRE_PREFIX) - 1;
const size_t name_len = sizeof("lov") - 1;
const size_t total_len = prefix_len + name_len + 1;
if (((rc + total_len) > size) && (buffer != NULL)) {
ptlrpc_req_finished(request);
return -ERANGE;
}
if (buffer != NULL) {
buffer += rc;
memcpy(buffer, XATTR_LUSTRE_PREFIX, prefix_len);
memcpy(buffer + prefix_len, "lov", name_len);
buffer[prefix_len + name_len] = '\0';
}
rc2 = total_len;
}
out:
ptlrpc_req_finished(request);
rc = rc + rc2;
return rc;
}
ssize_t ll_getxattr(struct dentry *dentry, const char *name,
void *buffer, size_t size)
{
struct inode *inode = dentry->d_inode;
LASSERT(inode);
LASSERT(name);
CDEBUG(D_VFSTRACE, "VFS Op:inode=%lu/%u(%p), xattr %s\n",
inode->i_ino, inode->i_generation, inode, name);
ll_stats_ops_tally(ll_i2sbi(inode), LPROC_LL_GETXATTR, 1);
if ((strncmp(name, XATTR_TRUSTED_PREFIX,
sizeof(XATTR_TRUSTED_PREFIX) - 1) == 0 &&
strcmp(name + sizeof(XATTR_TRUSTED_PREFIX) - 1, "lov") == 0) ||
(strncmp(name, XATTR_LUSTRE_PREFIX,
sizeof(XATTR_LUSTRE_PREFIX) - 1) == 0 &&
strcmp(name + sizeof(XATTR_LUSTRE_PREFIX) - 1, "lov") == 0)) {
struct lov_stripe_md *lsm;
struct lov_user_md *lump;
struct lov_mds_md *lmm = NULL;
struct ptlrpc_request *request = NULL;
int rc = 0, lmmsize = 0;
if (!S_ISREG(inode->i_mode) && !S_ISDIR(inode->i_mode))
return -ENODATA;
if (size == 0 && S_ISDIR(inode->i_mode)) {
/* XXX directory EA is fix for now, optimize to save
* RPC transfer */
GOTO(out, rc = sizeof(struct lov_user_md));
}
lsm = ccc_inode_lsm_get(inode);
if (lsm == NULL) {
if (S_ISDIR(inode->i_mode)) {
rc = ll_dir_getstripe(inode, &lmm,
&lmmsize, &request);
} else {
rc = -ENODATA;
}
} else {
/* LSM is present already after lookup/getattr call.
* we need to grab layout lock once it is implemented */
rc = obd_packmd(ll_i2dtexp(inode), &lmm, lsm);
lmmsize = rc;
}
ccc_inode_lsm_put(inode, lsm);
if (rc < 0)
GOTO(out, rc);
if (size == 0) {
/* used to call ll_get_max_mdsize() forward to get
* the maximum buffer size, while some apps (such as
* rsync 3.0.x) care much about the exact xattr value
* size */
rc = lmmsize;
GOTO(out, rc);
}
if (size < lmmsize) {
CERROR("server bug: replied size %d > %d for %s (%s)\n",
lmmsize, (int)size, dentry->d_name.name, name);
GOTO(out, rc = -ERANGE);
}
lump = (struct lov_user_md *)buffer;
memcpy(lump, lmm, lmmsize);
/* do not return layout gen for getxattr otherwise it would
* confuse tar --xattr by recognizing layout gen as stripe
* offset when the file is restored. See LU-2809. */
lump->lmm_layout_gen = 0;
rc = lmmsize;
out:
if (request)
ptlrpc_req_finished(request);
else if (lmm)
obd_free_diskmd(ll_i2dtexp(inode), &lmm);
return(rc);
}
return ll_getxattr_common(inode, name, buffer, size, OBD_MD_FLXATTR);
}
