static int gfs2_mmap(struct file *file, struct vm_area_struct *vma)
{
struct gfs2_inode *ip = GFS2_I(file->f_mapping->host);
if (!(file->f_flags & O_NOATIME) &&
!IS_NOATIME(&ip->i_inode)) {
struct gfs2_holder i_gh;
int error;
gfs2_holder_init(ip->i_gl, LM_ST_SHARED, LM_FLAG_ANY, &i_gh);
error = gfs2_glock_nq(&i_gh);
if (error == 0) {
file_accessed(file);
gfs2_glock_dq(&i_gh);
}
gfs2_holder_uninit(&i_gh);
if (error)
return error;
}
vma->vm_ops = &gfs2_vm_ops;
return 0;
}
static ssize_t hypfs_aio_read(struct kiocb *iocb, const struct iovec *iov,
unsigned long nr_segs, loff_t offset)
{
char *data;
ssize_t ret;
struct file *filp = iocb->ki_filp;
/* XXX: temporary */
char __user *buf = iov[0].iov_base;
size_t count = iov[0].iov_len;
if (nr_segs != 1)
return -EINVAL;
data = filp->private_data;
ret = simple_read_from_buffer(buf, count, &offset, data, strlen(data));
if (ret <= 0)
return ret;
iocb->ki_pos += ret;
file_accessed(filp);
return ret;
}
static ssize_t hypfs_aio_read(struct kiocb *iocb, __user char *buf,
size_t count, loff_t offset)
{
char *data;
size_t len;
struct file *filp = iocb->ki_filp;
data = filp->private_data;
len = strlen(data);
if (offset > len) {
count = 0;
goto out;
}
if (count > len - offset)
count = len - offset;
if (copy_to_user(buf, data + offset, count)) {
count = -EFAULT;
goto out;
}
iocb->ki_pos += count;
file_accessed(filp);
out:
return count;
}
/* file operations for directories */
static int coda_readdir(struct file *coda_file, struct dir_context *ctx)
{
struct coda_file_info *cfi;
struct file *host_file;
int ret;
cfi = CODA_FTOC(coda_file);
BUG_ON(!cfi || cfi->cfi_magic != CODA_MAGIC);
host_file = cfi->cfi_container;
if (host_file->f_op->iterate) {
struct inode *host_inode = file_inode(host_file);
mutex_lock(&host_inode->i_mutex);
ret = -ENOENT;
if (!IS_DEADDIR(host_inode)) {
ret = host_file->f_op->iterate(host_file, ctx);
file_accessed(host_file);
}
mutex_unlock(&host_inode->i_mutex);
return ret;
}
/* Venus: we must read Venus dirents from a file */
return coda_venus_readdir(coda_file, ctx);
}
int vfs_readdir(struct file *file, filldir_t filler, void *buf)
{
struct inode *inode = file->f_path.dentry->d_inode;
int res = -ENOTDIR;
if (!file->f_op || !file->f_op->readdir)
goto out;
res = security_file_permission(file, MAY_READ);
if (res)
goto out;
res = mutex_lock_killable(&inode->i_mutex);
if (res)
goto out;
res = -ENOENT;
if (!IS_DEADDIR(inode)) {
res = file->f_op->readdir(file, buf, filler);
file_accessed(file);
}
mutex_unlock(&inode->i_mutex);
out:
return res;
}
int f2fs_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
{
struct inode *inode = file->f_mapping->host;
struct f2fs_inode_info *fi = F2FS_I(inode);
struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
nid_t ino = inode->i_ino;
int ret = 0;
bool need_cp = false;
struct writeback_control wbc = {
.sync_mode = WB_SYNC_ALL,
.nr_to_write = LONG_MAX,
.for_reclaim = 0,
};
if (unlikely(f2fs_readonly(inode->i_sb)))
return 0;
trace_f2fs_sync_file_enter(inode);
/* if fdatasync is triggered, let's do in-place-update */
if (get_dirty_pages(inode) <= SM_I(sbi)->min_fsync_blocks)
set_inode_flag(fi, FI_NEED_IPU);
ret = filemap_write_and_wait_range(inode->i_mapping, start, end);
clear_inode_flag(fi, FI_NEED_IPU);
if (ret) {
trace_f2fs_sync_file_exit(inode, need_cp, datasync, ret);
return ret;
}
/* if the inode is dirty, let's recover all the time */
if (!datasync && is_inode_flag_set(fi, FI_DIRTY_INODE)) {
update_inode_page(inode);
goto go_write;
}
/*
* if there is no written data, don't waste time to write recovery info.
*/
if (!is_inode_flag_set(fi, FI_APPEND_WRITE) &&
!exist_written_data(sbi, ino, APPEND_INO)) {
/* it may call write_inode just prior to fsync */
if (need_inode_page_update(sbi, ino))
goto go_write;
if (is_inode_flag_set(fi, FI_UPDATE_WRITE) ||
exist_written_data(sbi, ino, UPDATE_INO))
goto flush_out;
goto out;
}
go_write:
/* guarantee free sections for fsync */
f2fs_balance_fs(sbi);
/*
* Both of fdatasync() and fsync() are able to be recovered from
* sudden-power-off.
*/
down_read(&fi->i_sem);
need_cp = need_do_checkpoint(inode);
up_read(&fi->i_sem);
if (need_cp) {
/* all the dirty node pages should be flushed for POR */
ret = f2fs_sync_fs(inode->i_sb, 1);
/*
* We've secured consistency through sync_fs. Following pino
* will be used only for fsynced inodes after checkpoint.
*/
try_to_fix_pino(inode);
clear_inode_flag(fi, FI_APPEND_WRITE);
clear_inode_flag(fi, FI_UPDATE_WRITE);
goto out;
}
sync_nodes:
sync_node_pages(sbi, ino, &wbc);
/* if cp_error was enabled, we should avoid infinite loop */
if (unlikely(f2fs_cp_error(sbi)))
goto out;
if (need_inode_block_update(sbi, ino)) {
mark_inode_dirty_sync(inode);
f2fs_write_inode(inode, NULL);
goto sync_nodes;
}
ret = wait_on_node_pages_writeback(sbi, ino);
if (ret)
goto out;
/* once recovery info is written, don't need to tack this */
remove_dirty_inode(sbi, ino, APPEND_INO);
clear_inode_flag(fi, FI_APPEND_WRITE);
flush_out:
remove_dirty_inode(sbi, ino, UPDATE_INO);
clear_inode_flag(fi, FI_UPDATE_WRITE);
ret = f2fs_issue_flush(sbi);
out:
trace_f2fs_sync_file_exit(inode, need_cp, datasync, ret);
f2fs_trace_ios(NULL, 1);
return ret;
}
static pgoff_t __get_first_dirty_index(struct address_space *mapping,
pgoff_t pgofs, int whence)
{
struct pagevec pvec;
int nr_pages;
if (whence != SEEK_DATA)
return 0;
/* find first dirty page index */
pagevec_init(&pvec, 0);
nr_pages = pagevec_lookup_tag(&pvec, mapping, &pgofs,
PAGECACHE_TAG_DIRTY, 1);
pgofs = nr_pages ? pvec.pages[0]->index : LONG_MAX;
pagevec_release(&pvec);
return pgofs;
}
static bool __found_offset(block_t blkaddr, pgoff_t dirty, pgoff_t pgofs,
int whence)
{
switch (whence) {
case SEEK_DATA:
if ((blkaddr == NEW_ADDR && dirty == pgofs) ||
(blkaddr != NEW_ADDR && blkaddr != NULL_ADDR))
return true;
break;
case SEEK_HOLE:
if (blkaddr == NULL_ADDR)
return true;
break;
}
return false;
}
static inline int unsigned_offsets(struct file *file)
{
return file->f_mode & FMODE_UNSIGNED_OFFSET;
}
static loff_t vfs_setpos(struct file *file, loff_t offset, loff_t maxsize)
{
if (offset < 0 && !unsigned_offsets(file))
return -EINVAL;
if (offset > maxsize)
return -EINVAL;
if (offset != file->f_pos) {
file->f_pos = offset;
file->f_version = 0;
}
return offset;
}
static loff_t f2fs_seek_block(struct file *file, loff_t offset, int whence)
{
struct inode *inode = file->f_mapping->host;
loff_t maxbytes = inode->i_sb->s_maxbytes;
struct dnode_of_data dn;
pgoff_t pgofs, end_offset, dirty;
loff_t data_ofs = offset;
loff_t isize;
int err = 0;
mutex_lock(&inode->i_mutex);
isize = i_size_read(inode);
if (offset >= isize)
goto fail;
/* handle inline data case */
if (f2fs_has_inline_data(inode) || f2fs_has_inline_dentry(inode)) {
if (whence == SEEK_HOLE)
data_ofs = isize;
goto found;
}
pgofs = (pgoff_t)(offset >> PAGE_CACHE_SHIFT);
dirty = __get_first_dirty_index(inode->i_mapping, pgofs, whence);
for (; data_ofs < isize; data_ofs = pgofs << PAGE_CACHE_SHIFT) {
set_new_dnode(&dn, inode, NULL, NULL, 0);
err = get_dnode_of_data(&dn, pgofs, LOOKUP_NODE_RA);
if (err && err != -ENOENT) {
goto fail;
} else if (err == -ENOENT) {
/* direct node does not exists */
if (whence == SEEK_DATA) {
pgofs = PGOFS_OF_NEXT_DNODE(pgofs,
F2FS_I(inode));
continue;
} else {
goto found;
}
}
end_offset = ADDRS_PER_PAGE(dn.node_page, F2FS_I(inode));
/* find data/hole in dnode block */
for (; dn.ofs_in_node < end_offset;
dn.ofs_in_node++, pgofs++,
data_ofs = (loff_t)pgofs << PAGE_CACHE_SHIFT) {
block_t blkaddr;
blkaddr = datablock_addr(dn.node_page, dn.ofs_in_node);
if (__found_offset(blkaddr, dirty, pgofs, whence)) {
f2fs_put_dnode(&dn);
goto found;
}
}
f2fs_put_dnode(&dn);
}
if (whence == SEEK_DATA)
goto fail;
found:
if (whence == SEEK_HOLE && data_ofs > isize)
data_ofs = isize;
mutex_unlock(&inode->i_mutex);
return vfs_setpos(file, data_ofs, maxbytes);
fail:
mutex_unlock(&inode->i_mutex);
return -ENXIO;
}
static loff_t f2fs_llseek(struct file *file, loff_t offset, int whence)
{
struct inode *inode = file->f_mapping->host;
loff_t maxbytes = inode->i_sb->s_maxbytes;
switch (whence) {
case SEEK_SET:
case SEEK_CUR:
case SEEK_END:
return generic_file_llseek_size(file, offset, whence,
maxbytes);
case SEEK_DATA:
case SEEK_HOLE:
if (offset < 0)
return -ENXIO;
return f2fs_seek_block(file, offset, whence);
}
return -EINVAL;
}
static int f2fs_file_mmap(struct file *file, struct vm_area_struct *vma)
{
struct inode *inode = file_inode(file);
if (f2fs_encrypted_inode(inode)) {
int err = f2fs_get_encryption_info(inode);
if (err)
return 0;
}
/* we don't need to use inline_data strictly */
if (f2fs_has_inline_data(inode)) {
int err = f2fs_convert_inline_inode(inode);
if (err)
return err;
}
file_accessed(file);
vma->vm_ops = &f2fs_file_vm_ops;
return 0;
}
static int f2fs_file_open(struct inode *inode, struct file *filp)
{
int ret = generic_file_open(inode, filp);
if (!ret && f2fs_encrypted_inode(inode)) {
ret = f2fs_get_encryption_info(inode);
if (ret)
ret = -EACCES;
}
return ret;
}
static ssize_t
pipe_readv(struct file *filp, const struct iovec *_iov,
unsigned long nr_segs, loff_t *ppos)
{
struct inode *inode = filp->f_dentry->d_inode;
int do_wakeup;
ssize_t ret;
struct iovec *iov = (struct iovec *)_iov;
size_t total_len;
total_len = iov_length(iov, nr_segs);
/* Null read succeeds. */
if (unlikely(total_len == 0))
return 0;
do_wakeup = 0;
ret = 0;
down(PIPE_SEM(*inode));
for (;;) {
int size = PIPE_LEN(*inode);
if (size) {
char *pipebuf = PIPE_BASE(*inode) + PIPE_START(*inode);
ssize_t chars = PIPE_MAX_RCHUNK(*inode);
if (chars > total_len)
chars = total_len;
if (chars > size)
chars = size;
if (pipe_iov_copy_to_user(iov, pipebuf, chars)) {
if (!ret) ret = -EFAULT;
break;
}
ret += chars;
PIPE_START(*inode) += chars;
PIPE_START(*inode) &= (PIPE_SIZE - 1);
PIPE_LEN(*inode) -= chars;
total_len -= chars;
do_wakeup = 1;
if (!total_len)
break; /* common path: read succeeded */
}
if (PIPE_LEN(*inode)) /* test for cyclic buffers */
continue;
if (!PIPE_WRITERS(*inode))
break;
if (!PIPE_WAITING_WRITERS(*inode)) {
/* syscall merging: Usually we must not sleep
* if O_NONBLOCK is set, or if we got some data.
* But if a writer sleeps in kernel space, then
* we can wait for that data without violating POSIX.
*/
if (ret)
break;
if (filp->f_flags & O_NONBLOCK) {
ret = -EAGAIN;
break;
}
}
if (signal_pending(current)) {
if (!ret) ret = -ERESTARTSYS;
break;
}
if (do_wakeup) {
wake_up_interruptible_sync(PIPE_WAIT(*inode));
kill_fasync(PIPE_FASYNC_WRITERS(*inode), SIGIO, POLL_OUT);
}
pipe_wait(inode);
}
up(PIPE_SEM(*inode));
/* Signal writers asynchronously that there is more room. */
if (do_wakeup) {
wake_up_interruptible(PIPE_WAIT(*inode));
kill_fasync(PIPE_FASYNC_WRITERS(*inode), SIGIO, POLL_OUT);
}
/*
* Hack: we turn off atime updates for -RT kernels.
* Who uses them on pipes anyway?
*/
#ifndef CONFIG_PREEMPT_RT
if (ret > 0)
file_accessed(filp);
#endif
return ret;
}
/*
* Common entry point for read/write/readv/writev
* This function will dispatch it to either the direct I/O
* or buffered I/O path depending on the mount options and/or
* augmented/extended metadata attached to the file.
* Note: File extended attributes override any mount options.
*/
static ssize_t do_readv_writev(enum ORANGEFS_io_type type, struct file *file,
loff_t *offset, struct iov_iter *iter)
{
struct inode *inode = file->f_mapping->host;
struct orangefs_inode_s *orangefs_inode = ORANGEFS_I(inode);
struct orangefs_khandle *handle = &orangefs_inode->refn.khandle;
size_t count = iov_iter_count(iter);
ssize_t total_count = 0;
ssize_t ret = -EINVAL;
gossip_debug(GOSSIP_FILE_DEBUG,
"%s-BEGIN(%pU): count(%d) after estimate_max_iovecs.\n",
__func__,
handle,
(int)count);
if (type == ORANGEFS_IO_WRITE) {
gossip_debug(GOSSIP_FILE_DEBUG,
"%s(%pU): proceeding with offset : %llu, "
"size %d\n",
__func__,
handle,
llu(*offset),
(int)count);
}
if (count == 0) {
ret = 0;
goto out;
}
while (iov_iter_count(iter)) {
size_t each_count = iov_iter_count(iter);
size_t amt_complete;
/* how much to transfer in this loop iteration */
if (each_count > orangefs_bufmap_size_query())
each_count = orangefs_bufmap_size_query();
gossip_debug(GOSSIP_FILE_DEBUG,
"%s(%pU): size of each_count(%d)\n",
__func__,
handle,
(int)each_count);
gossip_debug(GOSSIP_FILE_DEBUG,
"%s(%pU): BEFORE wait_for_io: offset is %d\n",
__func__,
handle,
(int)*offset);
ret = wait_for_direct_io(type, inode, offset, iter,
each_count, 0);
gossip_debug(GOSSIP_FILE_DEBUG,
"%s(%pU): return from wait_for_io:%d\n",
__func__,
handle,
(int)ret);
if (ret < 0)
goto out;
*offset += ret;
total_count += ret;
amt_complete = ret;
gossip_debug(GOSSIP_FILE_DEBUG,
"%s(%pU): AFTER wait_for_io: offset is %d\n",
__func__,
handle,
(int)*offset);
/*
* if we got a short I/O operations,
* fall out and return what we got so far
*/
if (amt_complete < each_count)
break;
} /*end while */
out:
if (total_count > 0)
ret = total_count;
if (ret > 0) {
if (type == ORANGEFS_IO_READ) {
file_accessed(file);
} else {
file_update_time(file);
/*
* Must invalidate to ensure write loop doesn't
* prevent kernel from reading updated
* attribute. Size probably changed because of
* the write, and other clients could update
* any other attribute.
*/
orangefs_inode->getattr_time = jiffies - 1;
}
}
gossip_debug(GOSSIP_FILE_DEBUG,
"%s(%pU): Value(%d) returned.\n",
__func__,
handle,
(int)ret);
return ret;
}
static int gfs2_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf)
{
struct page *page = vmf->page;
struct inode *inode = file_inode(vma->vm_file);
struct gfs2_inode *ip = GFS2_I(inode);
struct gfs2_sbd *sdp = GFS2_SB(inode);
struct gfs2_alloc_parms ap = { .aflags = 0, };
unsigned long last_index;
u64 pos = page->index << PAGE_CACHE_SHIFT;
unsigned int data_blocks, ind_blocks, rblocks;
struct gfs2_holder gh;
loff_t size;
int ret;
sb_start_pagefault(inode->i_sb);
/* Update file times before taking page lock */
file_update_time(vma->vm_file);
ret = get_write_access(inode);
if (ret)
goto out;
ret = gfs2_rs_alloc(ip);
if (ret)
goto out_write_access;
gfs2_size_hint(vma->vm_file, pos, PAGE_CACHE_SIZE);
gfs2_holder_init(ip->i_gl, LM_ST_EXCLUSIVE, 0, &gh);
ret = gfs2_glock_nq(&gh);
if (ret)
goto out_uninit;
set_bit(GLF_DIRTY, &ip->i_gl->gl_flags);
set_bit(GIF_SW_PAGED, &ip->i_flags);
if (!gfs2_write_alloc_required(ip, pos, PAGE_CACHE_SIZE)) {
lock_page(page);
if (!PageUptodate(page) || page->mapping != inode->i_mapping) {
ret = -EAGAIN;
unlock_page(page);
}
goto out_unlock;
}
ret = gfs2_rindex_update(sdp);
if (ret)
goto out_unlock;
ret = gfs2_quota_lock_check(ip);
if (ret)
goto out_unlock;
gfs2_write_calc_reserv(ip, PAGE_CACHE_SIZE, &data_blocks, &ind_blocks);
ap.target = data_blocks + ind_blocks;
ret = gfs2_inplace_reserve(ip, &ap);
if (ret)
goto out_quota_unlock;
rblocks = RES_DINODE + ind_blocks;
if (gfs2_is_jdata(ip))
rblocks += data_blocks ? data_blocks : 1;
if (ind_blocks || data_blocks) {
rblocks += RES_STATFS + RES_QUOTA;
rblocks += gfs2_rg_blocks(ip, data_blocks + ind_blocks);
}
ret = gfs2_trans_begin(sdp, rblocks, 0);
if (ret)
goto out_trans_fail;
lock_page(page);
ret = -EINVAL;
size = i_size_read(inode);
last_index = (size - 1) >> PAGE_CACHE_SHIFT;
/* Check page index against inode size */
if (size == 0 || (page->index > last_index))
goto out_trans_end;
ret = -EAGAIN;
/* If truncated, we must retry the operation, we may have raced
* with the glock demotion code.
*/
if (!PageUptodate(page) || page->mapping != inode->i_mapping)
goto out_trans_end;
/* Unstuff, if required, and allocate backing blocks for page */
ret = 0;
if (gfs2_is_stuffed(ip))
ret = gfs2_unstuff_dinode(ip, page);
if (ret == 0)
ret = gfs2_allocate_page_backing(page);
out_trans_end:
if (ret)
unlock_page(page);
gfs2_trans_end(sdp);
out_trans_fail:
gfs2_inplace_release(ip);
out_quota_unlock:
gfs2_quota_unlock(ip);
out_unlock:
gfs2_glock_dq(&gh);
out_uninit:
gfs2_holder_uninit(&gh);
if (ret == 0) {
set_page_dirty(page);
wait_for_stable_page(page);
}
out_write_access:
put_write_access(inode);
out:
sb_end_pagefault(inode->i_sb);
return block_page_mkwrite_return(ret);
}
static const struct vm_operations_struct gfs2_vm_ops = {
.fault = filemap_fault,
.map_pages = filemap_map_pages,
.page_mkwrite = gfs2_page_mkwrite,
.remap_pages = generic_file_remap_pages,
};
/**
* gfs2_mmap -
* @file: The file to map
* @vma: The VMA which described the mapping
*
* There is no need to get a lock here unless we should be updating
* atime. We ignore any locking errors since the only consequence is
* a missed atime update (which will just be deferred until later).
*
* Returns: 0
*/
static int gfs2_mmap(struct file *file, struct vm_area_struct *vma)
{
struct gfs2_inode *ip = GFS2_I(file->f_mapping->host);
if (!(file->f_flags & O_NOATIME) &&
!IS_NOATIME(&ip->i_inode)) {
struct gfs2_holder i_gh;
int error;
error = gfs2_glock_nq_init(ip->i_gl, LM_ST_SHARED, LM_FLAG_ANY,
&i_gh);
if (error)
return error;
/* grab lock to update inode */
gfs2_glock_dq_uninit(&i_gh);
file_accessed(file);
}
vma->vm_ops = &gfs2_vm_ops;
return 0;
}
/**
* gfs2_open_common - This is common to open and atomic_open
* @inode: The inode being opened
* @file: The file being opened
*
* This maybe called under a glock or not depending upon how it has
* been called. We must always be called under a glock for regular
* files, however. For other file types, it does not matter whether
* we hold the glock or not.
*
* Returns: Error code or 0 for success
*/
int gfs2_open_common(struct inode *inode, struct file *file)
{
struct gfs2_file *fp;
int ret;
if (S_ISREG(inode->i_mode)) {
ret = generic_file_open(inode, file);
if (ret)
return ret;
}
fp = kzalloc(sizeof(struct gfs2_file), GFP_NOFS);
if (!fp)
return -ENOMEM;
mutex_init(&fp->f_fl_mutex);
gfs2_assert_warn(GFS2_SB(inode), !file->private_data);
file->private_data = fp;
return 0;
}
/**
* gfs2_open - open a file
* @inode: the inode to open
* @file: the struct file for this opening
*
* After atomic_open, this function is only used for opening files
* which are already cached. We must still get the glock for regular
* files to ensure that we have the file size uptodate for the large
* file check which is in the common code. That is only an issue for
* regular files though.
*
* Returns: errno
*/
static int gfs2_open(struct inode *inode, struct file *file)
{
struct gfs2_inode *ip = GFS2_I(inode);
struct gfs2_holder i_gh;
int error;
bool need_unlock = false;
if (S_ISREG(ip->i_inode.i_mode)) {
error = gfs2_glock_nq_init(ip->i_gl, LM_ST_SHARED, LM_FLAG_ANY,
&i_gh);
if (error)
return error;
need_unlock = true;
}
error = gfs2_open_common(inode, file);
if (need_unlock)
gfs2_glock_dq_uninit(&i_gh);
return error;
}
/**
* gfs2_release - called to close a struct file
* @inode: the inode the struct file belongs to
* @file: the struct file being closed
*
* Returns: errno
*/
static int gfs2_release(struct inode *inode, struct file *file)
{
struct gfs2_inode *ip = GFS2_I(inode);
kfree(file->private_data);
file->private_data = NULL;
if (!(file->f_mode & FMODE_WRITE))
return 0;
gfs2_rs_delete(ip, &inode->i_writecount);
return 0;
}
/**
* gfs2_fsync - sync the dirty data for a file (across the cluster)
* @file: the file that points to the dentry
* @start: the start position in the file to sync
* @end: the end position in the file to sync
* @datasync: set if we can ignore timestamp changes
*
* We split the data flushing here so that we don't wait for the data
* until after we've also sent the metadata to disk. Note that for
* data=ordered, we will write & wait for the data at the log flush
* stage anyway, so this is unlikely to make much of a difference
* except in the data=writeback case.
*
* If the fdatawrite fails due to any reason except -EIO, we will
* continue the remainder of the fsync, although we'll still report
* the error at the end. This is to match filemap_write_and_wait_range()
* behaviour.
*
* Returns: errno
*/
static int gfs2_fsync(struct file *file, loff_t start, loff_t end,
int datasync)
{
struct address_space *mapping = file->f_mapping;
struct inode *inode = mapping->host;
int sync_state = inode->i_state & I_DIRTY;
struct gfs2_inode *ip = GFS2_I(inode);
int ret = 0, ret1 = 0;
if (mapping->nrpages) {
ret1 = filemap_fdatawrite_range(mapping, start, end);
if (ret1 == -EIO)
return ret1;
}
if (!gfs2_is_jdata(ip))
sync_state &= ~I_DIRTY_PAGES;
if (datasync)
sync_state &= ~I_DIRTY_SYNC;
if (sync_state) {
ret = sync_inode_metadata(inode, 1);
if (ret)
return ret;
if (gfs2_is_jdata(ip))
filemap_write_and_wait(mapping);
gfs2_ail_flush(ip->i_gl, 1);
}
if (mapping->nrpages)
ret = filemap_fdatawait_range(mapping, start, end);
return ret ? ret : ret1;
}
/**
* gfs2_file_aio_write - Perform a write to a file
* @iocb: The io context
* @iov: The data to write
* @nr_segs: Number of @iov segments
* @pos: The file position
*
* We have to do a lock/unlock here to refresh the inode size for
* O_APPEND writes, otherwise we can land up writing at the wrong
* offset. There is still a race, but provided the app is using its
* own file locking, this will make O_APPEND work as expected.
*
*/
static ssize_t gfs2_file_aio_write(struct kiocb *iocb, const struct iovec *iov,
unsigned long nr_segs, loff_t pos)
{
struct file *file = iocb->ki_filp;
size_t writesize = iov_length(iov, nr_segs);
struct gfs2_inode *ip = GFS2_I(file_inode(file));
int ret;
ret = gfs2_rs_alloc(ip);
if (ret)
return ret;
gfs2_size_hint(file, pos, writesize);
if (file->f_flags & O_APPEND) {
struct gfs2_holder gh;
ret = gfs2_glock_nq_init(ip->i_gl, LM_ST_SHARED, 0, &gh);
if (ret)
return ret;
gfs2_glock_dq_uninit(&gh);
}
return generic_file_aio_write(iocb, iov, nr_segs, pos);
}
static int fallocate_chunk(struct inode *inode, loff_t offset, loff_t len,
int mode)
{
struct gfs2_inode *ip = GFS2_I(inode);
struct buffer_head *dibh;
int error;
loff_t size = len;
unsigned int nr_blks;
sector_t lblock = offset >> inode->i_blkbits;
error = gfs2_meta_inode_buffer(ip, &dibh);
if (unlikely(error))
return error;
gfs2_trans_add_meta(ip->i_gl, dibh);
if (gfs2_is_stuffed(ip)) {
error = gfs2_unstuff_dinode(ip, NULL);
if (unlikely(error))
goto out;
}
while (len) {
struct buffer_head bh_map = { .b_state = 0, .b_blocknr = 0 };
bh_map.b_size = len;
set_buffer_zeronew(&bh_map);
error = gfs2_block_map(inode, lblock, &bh_map, 1);
if (unlikely(error))
goto out;
len -= bh_map.b_size;
nr_blks = bh_map.b_size >> inode->i_blkbits;
lblock += nr_blks;
if (!buffer_new(&bh_map))
continue;
if (unlikely(!buffer_zeronew(&bh_map))) {
error = -EIO;
goto out;
}
}
if (offset + size > inode->i_size && !(mode & FALLOC_FL_KEEP_SIZE))
i_size_write(inode, offset + size);
mark_inode_dirty(inode);
out:
brelse(dibh);
return error;
}
static void calc_max_reserv(struct gfs2_inode *ip, loff_t max, loff_t *len,
unsigned int *data_blocks, unsigned int *ind_blocks)
{
const struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
unsigned int max_blocks = ip->i_rgd->rd_free_clone;
unsigned int tmp, max_data = max_blocks - 3 * (sdp->sd_max_height - 1);
for (tmp = max_data; tmp > sdp->sd_diptrs;) {
tmp = DIV_ROUND_UP(tmp, sdp->sd_inptrs);
max_data -= tmp;
}
/* This calculation isn't the exact reverse of gfs2_write_calc_reserve,
so it might end up with fewer data blocks */
if (max_data <= *data_blocks)
return;
*data_blocks = max_data;
*ind_blocks = max_blocks - max_data;
*len = ((loff_t)max_data - 3) << sdp->sd_sb.sb_bsize_shift;
if (*len > max) {
*len = max;
gfs2_write_calc_reserv(ip, max, data_blocks, ind_blocks);
}
}
static long gfs2_fallocate(struct file *file, int mode, loff_t offset,
loff_t len)
{
struct inode *inode = file_inode(file);
struct gfs2_sbd *sdp = GFS2_SB(inode);
struct gfs2_inode *ip = GFS2_I(inode);
struct gfs2_alloc_parms ap = { .aflags = 0, };
unsigned int data_blocks = 0, ind_blocks = 0, rblocks;
loff_t bytes, max_bytes;
int error;
const loff_t pos = offset;
const loff_t count = len;
loff_t bsize_mask = ~((loff_t)sdp->sd_sb.sb_bsize - 1);
loff_t next = (offset + len - 1) >> sdp->sd_sb.sb_bsize_shift;
loff_t max_chunk_size = UINT_MAX & bsize_mask;
struct gfs2_holder gh;
next = (next + 1) << sdp->sd_sb.sb_bsize_shift;
/* We only support the FALLOC_FL_KEEP_SIZE mode */
if (mode & ~FALLOC_FL_KEEP_SIZE)
return -EOPNOTSUPP;
offset &= bsize_mask;
len = next - offset;
bytes = sdp->sd_max_rg_data * sdp->sd_sb.sb_bsize / 2;
if (!bytes)
bytes = UINT_MAX;
bytes &= bsize_mask;
if (bytes == 0)
bytes = sdp->sd_sb.sb_bsize;
error = gfs2_rs_alloc(ip);
if (error)
return error;
mutex_lock(&inode->i_mutex);
gfs2_holder_init(ip->i_gl, LM_ST_EXCLUSIVE, 0, &gh);
error = gfs2_glock_nq(&gh);
if (unlikely(error))
goto out_uninit;
gfs2_size_hint(file, offset, len);
while (len > 0) {
if (len < bytes)
bytes = len;
if (!gfs2_write_alloc_required(ip, offset, bytes)) {
len -= bytes;
offset += bytes;
continue;
}
error = gfs2_quota_lock_check(ip);
if (error)
goto out_unlock;
retry:
gfs2_write_calc_reserv(ip, bytes, &data_blocks, &ind_blocks);
ap.target = data_blocks + ind_blocks;
error = gfs2_inplace_reserve(ip, &ap);
if (error) {
if (error == -ENOSPC && bytes > sdp->sd_sb.sb_bsize) {
bytes >>= 1;
bytes &= bsize_mask;
if (bytes == 0)
bytes = sdp->sd_sb.sb_bsize;
goto retry;
}
goto out_qunlock;
}
max_bytes = bytes;
calc_max_reserv(ip, (len > max_chunk_size)? max_chunk_size: len,
&max_bytes, &data_blocks, &ind_blocks);
rblocks = RES_DINODE + ind_blocks + RES_STATFS + RES_QUOTA +
RES_RG_HDR + gfs2_rg_blocks(ip, data_blocks + ind_blocks);
if (gfs2_is_jdata(ip))
rblocks += data_blocks ? data_blocks : 1;
error = gfs2_trans_begin(sdp, rblocks,
PAGE_CACHE_SIZE/sdp->sd_sb.sb_bsize);
if (error)
goto out_trans_fail;
error = fallocate_chunk(inode, offset, max_bytes, mode);
gfs2_trans_end(sdp);
if (error)
goto out_trans_fail;
len -= max_bytes;
offset += max_bytes;
gfs2_inplace_release(ip);
gfs2_quota_unlock(ip);
}
if (error == 0)
error = generic_write_sync(file, pos, count);
goto out_unlock;
out_trans_fail:
gfs2_inplace_release(ip);
out_qunlock:
gfs2_quota_unlock(ip);
out_unlock:
gfs2_glock_dq(&gh);
out_uninit:
gfs2_holder_uninit(&gh);
mutex_unlock(&inode->i_mutex);
return error;
}
#ifdef CONFIG_GFS2_FS_LOCKING_DLM
/**
* gfs2_setlease - acquire/release a file lease
* @file: the file pointer
* @arg: lease type
* @fl: file lock
*
* We don't currently have a way to enforce a lease across the whole
* cluster; until we do, disable leases (by just returning -EINVAL),
* unless the administrator has requested purely local locking.
*
* Locking: called under i_lock
*
* Returns: errno
*/
static int gfs2_setlease(struct file *file, long arg, struct file_lock **fl)
{
return -EINVAL;
}
/**
* gfs2_lock - acquire/release a posix lock on a file
* @file: the file pointer
* @cmd: either modify or retrieve lock state, possibly wait
* @fl: type and range of lock
*
* Returns: errno
*/
static int gfs2_lock(struct file *file, int cmd, struct file_lock *fl)
{
struct gfs2_inode *ip = GFS2_I(file->f_mapping->host);
struct gfs2_sbd *sdp = GFS2_SB(file->f_mapping->host);
struct lm_lockstruct *ls = &sdp->sd_lockstruct;
if (!(fl->fl_flags & FL_POSIX))
return -ENOLCK;
if (__mandatory_lock(&ip->i_inode) && fl->fl_type != F_UNLCK)
return -ENOLCK;
if (cmd == F_CANCELLK) {
/* Hack: */
cmd = F_SETLK;
fl->fl_type = F_UNLCK;
}
if (unlikely(test_bit(SDF_SHUTDOWN, &sdp->sd_flags))) {
if (fl->fl_type == F_UNLCK)
posix_lock_file_wait(file, fl);
return -EIO;
}
if (IS_GETLK(cmd))
return dlm_posix_get(ls->ls_dlm, ip->i_no_addr, file, fl);
else if (fl->fl_type == F_UNLCK)
return dlm_posix_unlock(ls->ls_dlm, ip->i_no_addr, file, fl);
else
return dlm_posix_lock(ls->ls_dlm, ip->i_no_addr, file, cmd, fl);
}
static int do_flock(struct file *file, int cmd, struct file_lock *fl)
{
struct gfs2_file *fp = file->private_data;
struct gfs2_holder *fl_gh = &fp->f_fl_gh;
struct gfs2_inode *ip = GFS2_I(file_inode(file));
struct gfs2_glock *gl;
unsigned int state;
int flags;
int error = 0;
state = (fl->fl_type == F_WRLCK) ? LM_ST_EXCLUSIVE : LM_ST_SHARED;
flags = (IS_SETLKW(cmd) ? 0 : LM_FLAG_TRY) | GL_EXACT | GL_NOCACHE;
mutex_lock(&fp->f_fl_mutex);
gl = fl_gh->gh_gl;
if (gl) {
if (fl_gh->gh_state == state)
goto out;
flock_lock_file_wait(file,
&(struct file_lock){.fl_type = F_UNLCK});
gfs2_glock_dq_wait(fl_gh);
gfs2_holder_reinit(state, flags, fl_gh);
} else {
error = gfs2_glock_get(GFS2_SB(&ip->i_inode), ip->i_no_addr,
&gfs2_flock_glops, CREATE, &gl);
if (error)
goto out;
gfs2_holder_init(gl, state, flags, fl_gh);
gfs2_glock_put(gl);
}
error = gfs2_glock_nq(fl_gh);
if (error) {
gfs2_holder_uninit(fl_gh);
if (error == GLR_TRYFAILED)
error = -EAGAIN;
} else {
error = flock_lock_file_wait(file, fl);
gfs2_assert_warn(GFS2_SB(&ip->i_inode), !error);
}
out:
mutex_unlock(&fp->f_fl_mutex);
return error;
}
STATIC ssize_t
xfs_file_dio_aio_read(
struct kiocb *iocb,
struct iov_iter *to)
{
struct address_space *mapping = iocb->ki_filp->f_mapping;
struct inode *inode = mapping->host;
struct xfs_inode *ip = XFS_I(inode);
loff_t isize = i_size_read(inode);
size_t count = iov_iter_count(to);
struct iov_iter data;
struct xfs_buftarg *target;
ssize_t ret = 0;
trace_xfs_file_direct_read(ip, count, iocb->ki_pos);
if (!count)
return 0; /* skip atime */
if (XFS_IS_REALTIME_INODE(ip))
target = ip->i_mount->m_rtdev_targp;
else
target = ip->i_mount->m_ddev_targp;
/* DIO must be aligned to device logical sector size */
if ((iocb->ki_pos | count) & target->bt_logical_sectormask) {
if (iocb->ki_pos == isize)
return 0;
return -EINVAL;
}
file_accessed(iocb->ki_filp);
/*
* Locking is a bit tricky here. If we take an exclusive lock for direct
* IO, we effectively serialise all new concurrent read IO to this file
* and block it behind IO that is currently in progress because IO in
* progress holds the IO lock shared. We only need to hold the lock
* exclusive to blow away the page cache, so only take lock exclusively
* if the page cache needs invalidation. This allows the normal direct
* IO case of no page cache pages to proceeed concurrently without
* serialisation.
*/
xfs_rw_ilock(ip, XFS_IOLOCK_SHARED);
if (mapping->nrpages) {
xfs_rw_iunlock(ip, XFS_IOLOCK_SHARED);
xfs_rw_ilock(ip, XFS_IOLOCK_EXCL);
/*
* The generic dio code only flushes the range of the particular
* I/O. Because we take an exclusive lock here, this whole
* sequence is considerably more expensive for us. This has a
* noticeable performance impact for any file with cached pages,
* even when outside of the range of the particular I/O.
*
* Hence, amortize the cost of the lock against a full file
* flush and reduce the chances of repeated iolock cycles going
* forward.
*/
if (mapping->nrpages) {
ret = filemap_write_and_wait(mapping);
if (ret) {
xfs_rw_iunlock(ip, XFS_IOLOCK_EXCL);
return ret;
}
/*
* Invalidate whole pages. This can return an error if
* we fail to invalidate a page, but this should never
* happen on XFS. Warn if it does fail.
*/
ret = invalidate_inode_pages2(mapping);
WARN_ON_ONCE(ret);
ret = 0;
}
xfs_rw_ilock_demote(ip, XFS_IOLOCK_EXCL);
}
data = *to;
ret = __blockdev_direct_IO(iocb, inode, target->bt_bdev, &data,
xfs_get_blocks_direct, NULL, NULL, 0);
if (ret >= 0) {
iocb->ki_pos += ret;
iov_iter_advance(to, ret);
}
xfs_rw_iunlock(ip, XFS_IOLOCK_SHARED);
return ret;
}
static int vvp_io_read_start(const struct lu_env *env,
const struct cl_io_slice *ios)
{
struct vvp_io *vio = cl2vvp_io(env, ios);
struct ccc_io *cio = cl2ccc_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 ll_ra_read *bead = &vio->cui_bead;
struct file *file = cio->cui_fd->fd_file;
int result;
loff_t pos = io->u.ci_rd.rd.crw_pos;
long cnt = io->u.ci_rd.rd.crw_count;
long tot = cio->cui_tot_count;
int exceed = 0;
CLOBINVRNT(env, obj, ccc_object_invariant(obj));
CDEBUG(D_VFSTRACE, "read: -> [%lli, %lli)\n", pos, pos + cnt);
if (!can_populate_pages(env, io, inode))
return 0;
result = ccc_prep_size(env, obj, io, pos, tot, &exceed);
if (result != 0)
return result;
else if (exceed != 0)
goto out;
LU_OBJECT_HEADER(D_INODE, env, &obj->co_lu,
"Read ino %lu, %lu bytes, offset %lld, size %llu\n",
inode->i_ino, cnt, pos, i_size_read(inode));
/* turn off the kernel's read-ahead */
cio->cui_fd->fd_file->f_ra.ra_pages = 0;
/* initialize read-ahead window once per syscall */
if (!vio->cui_ra_window_set) {
vio->cui_ra_window_set = 1;
bead->lrr_start = cl_index(obj, pos);
bead->lrr_count = cl_index(obj, tot + PAGE_CACHE_SIZE - 1);
ll_ra_read_in(file, bead);
}
/* BUG: 5972 */
file_accessed(file);
switch (vio->cui_io_subtype) {
case IO_NORMAL:
result = lustre_generic_file_read(file, cio, &pos);
break;
case IO_SPLICE:
result = generic_file_splice_read(file, &pos,
vio->u.splice.cui_pipe, cnt,
vio->u.splice.cui_flags);
/* LU-1109: do splice read stripe by stripe otherwise if it
* may make nfsd stuck if this read occupied all internal pipe
* buffers. */
io->ci_continue = 0;
break;
default:
CERROR("Wrong IO type %u\n", vio->cui_io_subtype);
LBUG();
}
out:
if (result >= 0) {
if (result < cnt)
io->ci_continue = 0;
io->ci_nob += result;
ll_rw_stats_tally(ll_i2sbi(inode), current->pid, cio->cui_fd,
pos, result, READ);
result = 0;
}
return result;
}
static ssize_t
ncp_file_read_iter(struct kiocb *iocb, struct iov_iter *to)
{
struct file *file = iocb->ki_filp;
struct inode *inode = file_inode(file);
size_t already_read = 0;
off_t pos = iocb->ki_pos;
size_t bufsize;
int error;
void *freepage;
size_t freelen;
ncp_dbg(1, "enter %pD2\n", file);
if (!iov_iter_count(to))
return 0;
if (pos > inode->i_sb->s_maxbytes)
return 0;
iov_iter_truncate(to, inode->i_sb->s_maxbytes - pos);
error = ncp_make_open(inode, O_RDONLY);
if (error) {
ncp_dbg(1, "open failed, error=%d\n", error);
return error;
}
bufsize = NCP_SERVER(inode)->buffer_size;
error = -EIO;
freelen = ncp_read_bounce_size(bufsize);
freepage = vmalloc(freelen);
if (!freepage)
goto outrel;
error = 0;
/* First read in as much as possible for each bufsize. */
while (iov_iter_count(to)) {
int read_this_time;
size_t to_read = min_t(size_t,
bufsize - (pos % bufsize),
iov_iter_count(to));
error = ncp_read_bounce(NCP_SERVER(inode),
NCP_FINFO(inode)->file_handle,
pos, to_read, to, &read_this_time,
freepage, freelen);
if (error) {
error = -EIO; /* NW errno -> Linux errno */
break;
}
pos += read_this_time;
already_read += read_this_time;
if (read_this_time != to_read)
break;
}
vfree(freepage);
iocb->ki_pos = pos;
file_accessed(file);
ncp_dbg(1, "exit %pD2\n", file);
outrel:
ncp_inode_close(inode);
return already_read ? already_read : error;
}
static int wrapfs_mmap(struct file *file, struct vm_area_struct *vma)
{
int err = 0;
bool willwrite;
struct file *lower_file;
const struct vm_operations_struct *saved_vm_ops = NULL;
#ifdef DEBUG_SUPPORT
if(debug_support(wrapfs_lower_file(file)->f_dentry->d_sb,"file"))
UDBG;
#endif
/* this might be deferred to mmap's writepage */
willwrite = ((vma->vm_flags | VM_SHARED | VM_WRITE) == vma->vm_flags);
/*
* File systems which do not implement ->writepage may use
* generic_file_readonly_mmap as their ->mmap op. If you call
* generic_file_readonly_mmap with VM_WRITE, you'd get an -EINVAL.
* But we cannot call the lower ->mmap op, so we can't tell that
* writeable mappings won't work. Therefore, our only choice is to
* check if the lower file system supports the ->writepage, and if
* not, return EINVAL (the same error that
* generic_file_readonly_mmap returns in that case).
*/
lower_file = wrapfs_lower_file(file);
if (willwrite && !lower_file->f_mapping->a_ops->writepage) {
err = -EINVAL;
printk(KERN_ERR "wrapfs: lower file system does not "
"support writeable mmap\n");
goto out;
}
/*
* find and save lower vm_ops.
*
* XXX: the VFS should have a cleaner way of finding the lower vm_ops
*/
if (!WRAPFS_F(file)->lower_vm_ops) {
err = lower_file->f_op->mmap(lower_file, vma);
if (err) {
printk(KERN_ERR "wrapfs: lower mmap failed %d\n", err);
goto out;
}
saved_vm_ops = vma->vm_ops; /* save: came from lower ->mmap */
err = do_munmap(current->mm, vma->vm_start,
vma->vm_end - vma->vm_start);
if (err) {
printk(KERN_ERR "wrapfs: do_munmap failed %d\n", err);
goto out;
}
}
/*
* Next 3 lines are all I need from generic_file_mmap. I definitely
* don't want its test for ->readpage which returns -ENOEXEC.
*/
file_accessed(file);
vma->vm_ops = &wrapfs_vm_ops;
vma->vm_flags |= VM_CAN_NONLINEAR;
file->f_mapping->a_ops = &wrapfs_aops; /* set our aops */
if (!WRAPFS_F(file)->lower_vm_ops) /* save for our ->fault */
WRAPFS_F(file)->lower_vm_ops = saved_vm_ops;
out:
#ifdef DEBUG_SUPPORT
if(debug_support(lower_file->f_dentry->d_sb,"file"))
UDBGE(err);
#endif
return err;
}
static int nilfs_file_mmap(struct file *file, struct vm_area_struct *vma)
{
file_accessed(file);
vma->vm_ops = &nilfs_file_vm_ops;
return 0;
}
static int sdcardfs_mmap(struct file *file, struct vm_area_struct *vma)
{
int err = 0;
bool willwrite;
struct file *lower_file;
const struct vm_operations_struct *saved_vm_ops = NULL;
#ifdef CONFIG_TIMA_IOMMU_OPT
if (vma->vm_end - vma->vm_start) {
/* iommu optimization- needs to be turned ON from
* the tz side.
*/
cpu_v7_tima_iommu_opt(vma->vm_start, vma->vm_end, (unsigned long)vma->vm_mm->pgd);
}
#endif /* CONFIG_TIMA_IOMMU_OPT */
/* this might be deferred to mmap's writepage */
willwrite = ((vma->vm_flags | VM_SHARED | VM_WRITE) == vma->vm_flags);
/*
* File systems which do not implement ->writepage may use
* generic_file_readonly_mmap as their ->mmap op. If you call
* generic_file_readonly_mmap with VM_WRITE, you'd get an -EINVAL.
* But we cannot call the lower ->mmap op, so we can't tell that
* writeable mappings won't work. Therefore, our only choice is to
* check if the lower file system supports the ->writepage, and if
* not, return EINVAL (the same error that
* generic_file_readonly_mmap returns in that case).
*/
lower_file = sdcardfs_lower_file(file);
if (willwrite && !lower_file->f_mapping->a_ops->writepage) {
err = -EINVAL;
printk(KERN_ERR "sdcardfs: lower file system does not "
"support writeable mmap\n");
goto out;
}
/*
* find and save lower vm_ops.
*
* XXX: the VFS should have a cleaner way of finding the lower vm_ops
*/
if (!SDCARDFS_F(file)->lower_vm_ops) {
err = lower_file->f_op->mmap(lower_file, vma);
if (err) {
printk(KERN_ERR "sdcardfs: lower mmap failed %d\n", err);
goto out;
}
saved_vm_ops = vma->vm_ops; /* save: came from lower ->mmap */
err = do_munmap(current->mm, vma->vm_start,
vma->vm_end - vma->vm_start);
if (err) {
printk(KERN_ERR "sdcardfs: do_munmap failed %d\n", err);
goto out;
}
}
/*
* Next 3 lines are all I need from generic_file_mmap. I definitely
* don't want its test for ->readpage which returns -ENOEXEC.
*/
file_accessed(file);
vma->vm_ops = &sdcardfs_vm_ops;
file->f_mapping->a_ops = &sdcardfs_aops; /* set our aops */
if (!SDCARDFS_F(file)->lower_vm_ops) /* save for our ->fault */
SDCARDFS_F(file)->lower_vm_ops = saved_vm_ops;
out:
return err;
}
ssize_t hmfs_xip_file_read(struct file * filp, char __user * buf, size_t len,
loff_t * ppos)
{
/* from do_XIP_mapping_read */
struct inode *inode = filp->f_inode;
pgoff_t index, end_index;
unsigned long offset;
loff_t isize, pos;
size_t copied = 0, error = 0;
pos = *ppos;
index = pos >> HMFS_PAGE_SIZE_BITS;
offset = pos & ~HMFS_PAGE_MASK;
mutex_lock(&inode->i_mutex);
isize = i_size_read(inode);
if (!isize)
goto out;
end_index = (isize - 1) >> HMFS_PAGE_SIZE_BITS;
/*
* nr : read length for this loop
* offset : start inner-blk offset this loop
* index : start inner-file blk number this loop
* copied : read length so far
* FIXME: pending for write_lock? anything like access_ok()
*/
do {
unsigned long nr, left;
void *xip_mem[1];
int zero = 0;
int size;
/* nr is the maximum number of bytes to copy from this page */
nr = HMFS_PAGE_SIZE; //HMFS_SIZE
if (index >= end_index) {
if (index > end_index)
goto out;
nr = ((isize - 1) & ~HMFS_PAGE_MASK) + 1;
if (nr <= offset) {
goto out;
}
}
nr = nr - offset;
if (nr > len - copied)
nr = len - copied;
BUG_ON(nr > HMFS_PAGE_SIZE);
error = get_data_blocks(inode, index, index + 1, xip_mem,
&size, RA_END);
if (unlikely(error || size != 1)) {
if (error == -ENODATA) {
/* sparse */
zero = 1;
} else
goto out;
}
/* copy to user space */
if (!zero)
left =
__copy_to_user(buf + copied, xip_mem[0] + offset, nr);
else
left = __clear_user(buf + copied, nr);
if (left) {
error = -EFAULT;
goto out;
}
copied += (nr - left);
offset += (nr - left);
index += offset >> HMFS_PAGE_SIZE_BITS;
offset &= ~HMFS_PAGE_MASK;
} while (copied < len);
out:
mutex_unlock(&inode->i_mutex);
*ppos = pos + copied;
if (filp)
file_accessed(filp);
return (copied ? copied : error);
}
ssize_t rawfs_reg_file_aio_write(struct kiocb *iocb, const struct iovec *iov,
unsigned long nr_segs, loff_t pos)
{
struct file *filp = iocb->ki_filp;
struct super_block *sb = filp->f_path.dentry->d_sb;
struct rawfs_sb_info *rawfs_sb = RAWFS_SB(sb);
// struct address_space *mapping=filp->f_mapping;
struct inode *inode = filp->f_mapping->host;
struct rawfs_inode_info *inode_info = RAWFS_I(inode);
ssize_t retval;
// size_t count;
loff_t *ppos = &iocb->ki_pos;
struct rawfs_file_info *fi = NULL;
unsigned int curr_file_pos = pos;
unsigned int curr_buf_pos = 0;
int starting_page, total_pages;
int remain_buf_size;
int result;
struct rawfs_file_list_entry * entry;
const struct iovec *iv = &iov[0]; // TODO: Process all io vectors
// Always use direct I/O
// loff_t size;
// int block_no;
// Get Lock
mutex_lock(&rawfs_sb->rawfs_lock);
retval=iov_length(iov, nr_segs);
RAWFS_PRINT(RAWFS_DBG_FILE, "rawfs_reg_file_aio_write %s, pos %lld, "
"len %d\n", inode_info->i_name, pos, retval);
fi = kzalloc(sizeof(struct rawfs_file_info), GFP_NOFS);
if (!fi) {
retval = 0;
goto out;
}
/* rawfs_fill_file_info(inode , fi); */
rawfs_fill_fileinfo_by_dentry(filp->f_path.dentry, fi);
/* Update file_info */
if ((pos+retval) > fi->i_size)
fi->i_size = pos+retval;
fi->i_chunk_index = 1;
fi->i_chunk_total = S_ISDIR(fi->i_mode) ? 1 : CEILING((unsigned)fi->i_size,
rawfs_sb->page_data_size);
/* do GC, if the required space is not enough */
result = rawfs_reserve_space(sb, fi->i_chunk_total);
if (result<0) {
retval=result;
goto out;
}
// get entry from file list
entry = rawfs_file_list_get(sb, fi->i_name, fi->i_parent_folder_id);
if (!entry) {
RAWFS_PRINT(RAWFS_DBG_FILE, "rawfs_reg_file_aio_write: %s file list "
"entry missing\n", fi->i_name);
dump_stack();
goto out;
}
// Write File
{
int preceding_pages, rear_pages;
struct rawfs_page *page_buf = NULL;
int i;
// Prepare page buffer
page_buf = kzalloc(rawfs_sb->page_size, GFP_NOFS);
if (page_buf == NULL) {
retval = 0;
goto out;
}
starting_page = rawfs_sb->data_block_free_page_index;
preceding_pages = FLOOR((unsigned)pos, rawfs_sb->page_data_size);
total_pages = CEILING((unsigned)fi->i_size, rawfs_sb->page_data_size);
rear_pages = CEILING((unsigned)pos + iv->iov_len,
rawfs_sb->page_data_size);
remain_buf_size = iv->iov_len;
RAWFS_PRINT(RAWFS_DBG_FILE, "rawfs_reg_file_aio_write %s, "
"preceding_pages %d, rear_pages %d, total_pages %d, "
"remain_buf_size %d\n",
inode_info->i_name, preceding_pages, rear_pages, total_pages,
remain_buf_size);
// Step 1: Copy preceding pages, if starting pos is not 0.
for (i=0;i<total_pages;i++)
{
// Read, if necessary, (no need for new files)
if ( (i<=preceding_pages) || (i>=rear_pages))
rawfs_sb->dev.read_page(sb,
entry->i_location_block,
entry->i_location_page+i,
page_buf);
// Update info
memcpy(&page_buf->i_info.i_file_info, fi,
sizeof(struct rawfs_file_info));
// Within Modify Range: Copy Modify
if ((i>=preceding_pages) && (i<=rear_pages))
{
int start_in_buf;
int copy_len;
start_in_buf = (curr_file_pos % rawfs_sb->page_data_size);
copy_len = ((start_in_buf + remain_buf_size) >
rawfs_sb->page_data_size) ?
(rawfs_sb->page_data_size - start_in_buf) :
remain_buf_size;
if (copy_from_user(&page_buf->i_data[0] + start_in_buf,
(char*)iv->iov_base + curr_buf_pos, copy_len))
{
retval = -EFAULT;
goto out2;
}
curr_buf_pos += copy_len;
remain_buf_size -= copy_len;
RAWFS_PRINT(RAWFS_DBG_FILE, "rawfs_reg_file_aio_write %s, %d, "
"curr_buf_pos %d, remain_buf_size %d start_in_buf %d "
"copy_len %d starting pattern %X\n",
inode_info->i_name, i, curr_buf_pos, remain_buf_size,
start_in_buf, copy_len,
*(unsigned int*)(&page_buf->i_data[0] + start_in_buf));
}
rawfs_page_signature(sb, page_buf);
// Write
rawfs_sb->dev.write_page(sb,
rawfs_sb->data_block,
rawfs_sb->data_block_free_page_index,
page_buf);
rawfs_sb->data_block_free_page_index++;
fi->i_chunk_index++;
}
out2:
if (page_buf)
kfree(page_buf);
}
if (retval > 0) {
*ppos = pos + retval;
}
file_accessed(filp);
// Update Inode: file size, block, page
i_size_write(inode, fi->i_size);
// TODO: Get inode lock when update
inode_info->i_location_block = rawfs_sb->data_block;
inode_info->i_location_page = starting_page;
inode_info->i_location_page_count = total_pages;
// update location
entry->i_location_block = rawfs_sb->data_block;
entry->i_location_page = starting_page;
entry->i_location_page_count = total_pages;
out:
if (fi)
kfree(fi);
// Release Lock
mutex_unlock(&rawfs_sb->rawfs_lock);
return retval;
}
//-----------------------------------------------------------------------------
// Regular File Operation
//-----------------------------------------------------------------------------
ssize_t rawfs_reg_file_aio_read(struct kiocb *iocb, const struct iovec *iov,
unsigned long nr_segs, loff_t pos)
{
struct file *filp = iocb->ki_filp;
struct super_block *sb = filp->f_path.dentry->d_sb;
struct rawfs_sb_info *rawfs_sb = RAWFS_SB(sb);
// struct address_space *mapping=filp->f_mapping;
struct inode *inode = filp->f_mapping->host;
struct rawfs_inode_info *inode_info = RAWFS_I(inode);
ssize_t retval;
size_t count;
loff_t *ppos = &iocb->ki_pos;
loff_t size;
unsigned int curr_file_pos = pos;
unsigned int curr_buf_pos = 0;
int remain_buf_size;
const struct iovec *iv = &iov[0]; // TODO: Process all io vectors
retval = generic_segment_checks(iov, &nr_segs, &count, VERIFY_WRITE);
RAWFS_PRINT(RAWFS_DBG_FILE, "rawfs_reg_file_aio_read %s, segment check "
"result %d\n", inode_info->i_name, retval);
if (retval)
return retval;
mutex_lock(&rawfs_sb->rawfs_lock);
retval=iov_length(iov, nr_segs);
size = i_size_read(inode);
RAWFS_PRINT(RAWFS_DBG_FILE, "rawfs_reg_file_aio_read %s, pos %lld, len %d, "
"filesize: %lld\n", inode_info->i_name, pos, retval, size);
if ((retval + pos) >= size)
retval = size - pos;
if (pos < size) {
/* Read File */
{
int preceding_pages, rear_pages;
struct rawfs_page *page_buf = NULL;
int i;
// Prepare page buffer
page_buf = kzalloc(rawfs_sb->page_size, GFP_NOFS);
if (page_buf == NULL) {
retval = 0;
goto out;
}
preceding_pages = FLOOR((unsigned)pos, rawfs_sb->page_data_size);
rear_pages = CEILING((unsigned)pos + retval,
rawfs_sb->page_data_size);
remain_buf_size = retval;
RAWFS_PRINT(RAWFS_DBG_FILE, "rawfs_reg_file_aio_read %s, "
"preceding_pages %d, rear_pages %d, remain_buf_size %d\n",
inode_info->i_name, preceding_pages, rear_pages,
remain_buf_size);
// Step 1: Copy preceding pages, if starting pos is not 0.
for (i=preceding_pages;i<rear_pages;i++)
{
__u32 crc;
// Read page
rawfs_sb->dev.read_page(sb,
inode_info->i_location_block,
inode_info->i_location_page+i,
page_buf);
// TODO: skip this page, if unrecoverable error occurs
/* CRC error should not happen,
since we have already check them at bootup */
crc = rawfs_page_crc_data(sb, page_buf);
if (crc != page_buf->i_crc)
RAWFS_PRINT(RAWFS_DBG_FILE, "rawfs_reg_file_aio_read: "
"%s @ %X, crc fail %X, expected %X\n",
page_buf->i_info.i_file_info.i_name,
page_buf->i_info.i_file_info.i_parent_folder_id, crc,
page_buf->i_crc);
else
RAWFS_PRINT(RAWFS_DBG_FILE, "rawfs_reg_file_aio_read: "
"%s @ %X, crc %X\n",
page_buf->i_info.i_file_info.i_name,
page_buf->i_info.i_file_info.i_parent_folder_id,
page_buf->i_crc);
/* Copy requried parts */
{
int start_in_buf;
int copy_len;
start_in_buf = (curr_file_pos % rawfs_sb->page_data_size);
copy_len = ((start_in_buf + remain_buf_size) >
rawfs_sb->page_data_size) ?
(rawfs_sb->page_data_size - start_in_buf) :
remain_buf_size;
if (copy_to_user((char*)iv->iov_base + curr_buf_pos,
&page_buf->i_data[0] + start_in_buf, copy_len))
{
retval = -EFAULT;
goto out2;
}
RAWFS_PRINT(RAWFS_DBG_FILE, "rawfs_reg_file_aio_read %s, "
"%d, curr_buf_pos %d, remain_buf_size %d "
"start_in_buf %d copy_len %d starting pattern %X\n",
inode_info->i_name, i, curr_buf_pos, remain_buf_size,
start_in_buf, copy_len,
*(unsigned int*)(&page_buf->i_data[0] + start_in_buf));
curr_buf_pos += copy_len;
remain_buf_size -= copy_len;
}
}
out2:
if (page_buf)
kfree(page_buf);
}
if (retval > 0)
*ppos = pos + retval;
if (retval < 0 || *ppos >= size) {
file_accessed(filp);
goto out;
}
}
else
retval = 0;
out:
// Release Lock
mutex_unlock(&rawfs_sb->rawfs_lock);
return retval;
}
/*
* Common entry point for read/write/readv/writev
* This function will dispatch it to either the direct I/O
* or buffered I/O path depending on the mount options and/or
* augmented/extended metadata attached to the file.
* Note: File extended attributes override any mount options.
*/
static ssize_t do_readv_writev(enum ORANGEFS_io_type type, struct file *file,
loff_t *offset, struct iov_iter *iter)
{
struct inode *inode = file->f_mapping->host;
struct orangefs_inode_s *orangefs_inode = ORANGEFS_I(inode);
struct orangefs_khandle *handle = &orangefs_inode->refn.khandle;
size_t count = iov_iter_count(iter);
ssize_t total_count = 0;
ssize_t ret = -EINVAL;
gossip_debug(GOSSIP_FILE_DEBUG,
"%s-BEGIN(%pU): count(%d) after estimate_max_iovecs.\n",
__func__,
handle,
(int)count);
if (type == ORANGEFS_IO_WRITE) {
gossip_debug(GOSSIP_FILE_DEBUG,
"%s(%pU): proceeding with offset : %llu, "
"size %d\n",
__func__,
handle,
llu(*offset),
(int)count);
}
if (count == 0) {
ret = 0;
goto out;
}
while (iov_iter_count(iter)) {
size_t each_count = iov_iter_count(iter);
size_t amt_complete;
/* how much to transfer in this loop iteration */
if (each_count > orangefs_bufmap_size_query())
each_count = orangefs_bufmap_size_query();
gossip_debug(GOSSIP_FILE_DEBUG,
"%s(%pU): size of each_count(%d)\n",
__func__,
handle,
(int)each_count);
gossip_debug(GOSSIP_FILE_DEBUG,
"%s(%pU): BEFORE wait_for_io: offset is %d\n",
__func__,
handle,
(int)*offset);
ret = wait_for_direct_io(type, inode, offset, iter,
each_count, 0);
gossip_debug(GOSSIP_FILE_DEBUG,
"%s(%pU): return from wait_for_io:%d\n",
__func__,
handle,
(int)ret);
if (ret < 0)
goto out;
*offset += ret;
total_count += ret;
amt_complete = ret;
gossip_debug(GOSSIP_FILE_DEBUG,
"%s(%pU): AFTER wait_for_io: offset is %d\n",
__func__,
handle,
(int)*offset);
/*
* if we got a short I/O operations,
* fall out and return what we got so far
*/
if (amt_complete < each_count)
break;
} /*end while */
out:
if (total_count > 0)
ret = total_count;
if (ret > 0) {
if (type == ORANGEFS_IO_READ) {
file_accessed(file);
} else {
SetMtimeFlag(orangefs_inode);
inode->i_mtime = CURRENT_TIME;
mark_inode_dirty_sync(inode);
}
}
gossip_debug(GOSSIP_FILE_DEBUG,
"%s(%pU): Value(%d) returned.\n",
__func__,
handle,
(int)ret);
return ret;
}
static ssize_t
pipe_read(struct kiocb *iocb, struct iov_iter *to)
{
size_t total_len = iov_iter_count(to);
struct file *filp = iocb->ki_filp;
struct pipe_inode_info *pipe = filp->private_data;
int do_wakeup;
ssize_t ret;
/* Null read succeeds. */
if (unlikely(total_len == 0))
return 0;
do_wakeup = 0;
ret = 0;
__pipe_lock(pipe);
for (;;) {
int bufs = pipe->nrbufs;
if (bufs) {
int curbuf = pipe->curbuf;
struct pipe_buffer *buf = pipe->bufs + curbuf;
const struct pipe_buf_operations *ops = buf->ops;
size_t chars = buf->len;
size_t written;
int error;
if (chars > total_len)
chars = total_len;
error = ops->confirm(pipe, buf);
if (error) {
if (!ret)
ret = error;
break;
}
written = copy_page_to_iter(buf->page, buf->offset, chars, to);
if (unlikely(written < chars)) {
if (!ret)
ret = -EFAULT;
break;
}
ret += chars;
buf->offset += chars;
buf->len -= chars;
/* Was it a packet buffer? Clean up and exit */
if (buf->flags & PIPE_BUF_FLAG_PACKET) {
total_len = chars;
buf->len = 0;
}
if (!buf->len) {
buf->ops = NULL;
ops->release(pipe, buf);
curbuf = (curbuf + 1) & (pipe->buffers - 1);
pipe->curbuf = curbuf;
pipe->nrbufs = --bufs;
do_wakeup = 1;
}
total_len -= chars;
if (!total_len)
break; /* common path: read succeeded */
}
if (bufs) /* More to do? */
continue;
if (!atomic_read(&pipe->writers))
break;
if (!atomic_read(&pipe->waiting_writers)) {
/* syscall merging: Usually we must not sleep
* if O_NONBLOCK is set, or if we got some data.
* But if a writer sleeps in kernel space, then
* we can wait for that data without violating POSIX.
*/
if (ret)
break;
if (filp->f_flags & O_NONBLOCK) {
ret = -EAGAIN;
break;
}
}
if (signal_pending(current)) {
if (!ret)
ret = -ERESTARTSYS;
break;
}
if (do_wakeup) {
wake_up_interruptible_sync_poll(&pipe->wait, POLLOUT | POLLWRNORM);
kill_fasync(&pipe->fasync_writers, SIGIO, POLL_OUT);
}
pipe_wait(pipe);
}
__pipe_unlock(pipe);
/* Signal writers asynchronously that there is more room. */
if (do_wakeup) {
wake_up_interruptible_sync_poll(&pipe->wait, POLLOUT | POLLWRNORM);
kill_fasync(&pipe->fasync_writers, SIGIO, POLL_OUT);
}
if (ret > 0)
file_accessed(filp);
return ret;
}
static int gfs2_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf)
{
struct page *page = vmf->page;
struct inode *inode = vma->vm_file->f_path.dentry->d_inode;
struct gfs2_inode *ip = GFS2_I(inode);
struct gfs2_sbd *sdp = GFS2_SB(inode);
struct gfs2_alloc_parms ap = { .aflags = 0, };
unsigned long last_index;
u64 pos = page->index << PAGE_CACHE_SHIFT;
unsigned int data_blocks, ind_blocks, rblocks;
int alloc_required = 0;
struct gfs2_holder gh;
loff_t size;
int ret;
sb_start_pagefault(inode->i_sb);
/* Update file times before taking page lock */
file_update_time(vma->vm_file);
ret = get_write_access(inode);
if (ret)
goto out;
ret = gfs2_rs_alloc(ip);
if (ret)
goto out_write_access;
gfs2_holder_init(ip->i_gl, LM_ST_EXCLUSIVE, 0, &gh);
ret = gfs2_glock_nq(&gh);
if (ret)
goto out_uninit;
set_bit(GLF_DIRTY, &ip->i_gl->gl_flags);
set_bit(GIF_SW_PAGED, &ip->i_flags);
gfs2_size_hint(inode, pos, PAGE_CACHE_SIZE);
ret = gfs2_write_alloc_required(ip, pos, PAGE_CACHE_SIZE, &alloc_required);
if (ret)
goto out_unlock;
if (!alloc_required) {
lock_page(page);
if (!PageUptodate(page) || page->mapping != inode->i_mapping) {
ret = -EAGAIN;
unlock_page(page);
}
goto out_unlock;
}
ret = gfs2_rindex_update(sdp);
if (ret)
goto out_unlock;
ret = gfs2_quota_lock_check(ip);
if (ret)
goto out_unlock;
gfs2_write_calc_reserv(ip, PAGE_CACHE_SIZE, &data_blocks, &ind_blocks);
ap.target = data_blocks + ind_blocks;
ret = gfs2_inplace_reserve(ip, &ap);
if (ret)
goto out_quota_unlock;
rblocks = RES_DINODE + ind_blocks;
if (gfs2_is_jdata(ip))
rblocks += data_blocks ? data_blocks : 1;
if (ind_blocks || data_blocks) {
rblocks += RES_STATFS + RES_QUOTA;
rblocks += gfs2_rg_blocks(ip, data_blocks + ind_blocks);
}
ret = gfs2_trans_begin(sdp, rblocks, 0);
if (ret)
goto out_trans_fail;
lock_page(page);
ret = -EINVAL;
size = i_size_read(inode);
last_index = (size - 1) >> PAGE_CACHE_SHIFT;
/* Check page index against inode size */
if (size == 0 || (page->index > last_index))
goto out_trans_end;
ret = -EAGAIN;
/* If truncated, we must retry the operation, we may have raced
* with the glock demotion code.
*/
if (!PageUptodate(page) || page->mapping != inode->i_mapping)
goto out_trans_end;
/* Unstuff, if required, and allocate backing blocks for page */
ret = 0;
if (gfs2_is_stuffed(ip))
ret = gfs2_unstuff_dinode(ip, page);
if (ret == 0)
ret = gfs2_allocate_page_backing(page);
out_trans_end:
if (ret)
unlock_page(page);
gfs2_trans_end(sdp);
out_trans_fail:
gfs2_inplace_release(ip);
out_quota_unlock:
gfs2_quota_unlock(ip);
out_unlock:
gfs2_glock_dq(&gh);
out_uninit:
gfs2_holder_uninit(&gh);
if (ret == 0) {
set_page_dirty(page);
wait_for_stable_page(page);
}
out_write_access:
put_write_access(inode);
out:
sb_end_pagefault(inode->i_sb);
return block_page_mkwrite_return(ret);
}
static const struct vm_operations_struct gfs2_vm_ops = {
.fault = filemap_fault,
.page_mkwrite = gfs2_page_mkwrite,
};
/**
* gfs2_mmap -
* @file: The file to map
* @vma: The VMA which described the mapping
*
* There is no need to get a lock here unless we should be updating
* atime. We ignore any locking errors since the only consequence is
* a missed atime update (which will just be deferred until later).
*
* Returns: 0
*/
static int gfs2_mmap(struct file *file, struct vm_area_struct *vma)
{
struct gfs2_inode *ip = GFS2_I(file->f_mapping->host);
if (!(file->f_flags & O_NOATIME) &&
!IS_NOATIME(&ip->i_inode)) {
struct gfs2_holder i_gh;
int error;
error = gfs2_glock_nq_init(ip->i_gl, LM_ST_SHARED, LM_FLAG_ANY,
&i_gh);
if (error)
return error;
/* grab lock to update inode */
gfs2_glock_dq_uninit(&i_gh);
file_accessed(file);
}
vma->vm_ops = &gfs2_vm_ops;
vma->vm_flags |= VM_CAN_NONLINEAR;
return 0;
}
/**
* gfs2_open - open a file
* @inode: the inode to open
* @file: the struct file for this opening
*
* Returns: errno
*/
static int gfs2_open(struct inode *inode, struct file *file)
{
struct gfs2_inode *ip = GFS2_I(inode);
struct gfs2_holder i_gh;
struct gfs2_file *fp;
int error;
fp = kzalloc(sizeof(struct gfs2_file), GFP_KERNEL);
if (!fp)
return -ENOMEM;
mutex_init(&fp->f_fl_mutex);
gfs2_assert_warn(GFS2_SB(inode), !file->private_data);
file->private_data = fp;
if (S_ISREG(ip->i_inode.i_mode)) {
error = gfs2_glock_nq_init(ip->i_gl, LM_ST_SHARED, LM_FLAG_ANY,
&i_gh);
if (error)
goto fail;
if (!(file->f_flags & O_LARGEFILE) &&
i_size_read(inode) > MAX_NON_LFS) {
error = -EOVERFLOW;
goto fail_gunlock;
}
gfs2_glock_dq_uninit(&i_gh);
}
return 0;
fail_gunlock:
gfs2_glock_dq_uninit(&i_gh);
fail:
file->private_data = NULL;
kfree(fp);
return error;
}
/**
* gfs2_release - called to close a struct file
* @inode: the inode the struct file belongs to
* @file: the struct file being closed
*
* Returns: errno
*/
static int gfs2_release(struct inode *inode, struct file *file)
{
struct gfs2_inode *ip = GFS2_I(inode);
kfree(file->private_data);
file->private_data = NULL;
if (!(file->f_mode & FMODE_WRITE))
return 0;
gfs2_rs_delete(ip);
return 0;
}
/**
* gfs2_fsync - sync the dirty data for a file (across the cluster)
* @file: the file that points to the dentry
* @start: the start position in the file to sync
* @end: the end position in the file to sync
* @datasync: set if we can ignore timestamp changes
*
* The VFS will flush data for us. We only need to worry
* about metadata here.
*
* Returns: errno
*/
static int gfs2_fsync(struct file *file, struct dentry *dentry, int datasync)
{
struct inode *inode = dentry->d_inode;
int sync_state = inode->i_state & I_DIRTY;
struct gfs2_inode *ip = GFS2_I(inode);
int ret;
if (!gfs2_is_jdata(ip))
sync_state &= ~I_DIRTY_PAGES;
if (datasync)
sync_state &= ~I_DIRTY_SYNC;
if (sync_state) {
ret = sync_inode_metadata(inode, 1);
if (ret)
return ret;
if (gfs2_is_jdata(ip))
filemap_write_and_wait(inode->i_mapping);
gfs2_ail_flush(ip->i_gl, 1);
}
return 0;
}
/**
* gfs2_file_aio_write - Perform a write to a file
* @iocb: The io context
* @iov: The data to write
* @nr_segs: Number of @iov segments
* @pos: The file position
*
* We have to do a lock/unlock here to refresh the inode size for
* O_APPEND writes, otherwise we can land up writing at the wrong
* offset. There is still a race, but provided the app is using its
* own file locking, this will make O_APPEND work as expected.
*
*/
static ssize_t gfs2_file_aio_write(struct kiocb *iocb, const struct iovec *iov,
unsigned long nr_segs, loff_t pos)
{
struct file *file = iocb->ki_filp;
size_t writesize = iov_length(iov, nr_segs);
struct dentry *dentry = file->f_dentry;
struct gfs2_inode *ip = GFS2_I(dentry->d_inode);
struct gfs2_sbd *sdp;
int ret;
sdp = GFS2_SB(file->f_mapping->host);
ret = gfs2_rs_alloc(ip);
if (ret)
return ret;
gfs2_size_hint(file->f_dentry->d_inode, pos, writesize);
if (file->f_flags & O_APPEND) {
struct gfs2_holder gh;
ret = gfs2_glock_nq_init(ip->i_gl, LM_ST_SHARED, 0, &gh);
if (ret)
return ret;
gfs2_glock_dq_uninit(&gh);
}
return generic_file_aio_write(iocb, iov, nr_segs, pos);
}
static ssize_t gfs2_file_splice_write(struct pipe_inode_info *pipe,
struct file *out, loff_t *ppos,
size_t len, unsigned int flags)
{
int error;
struct inode *inode = out->f_mapping->host;
struct gfs2_inode *ip = GFS2_I(inode);
error = gfs2_rs_alloc(ip);
if (error)
return (ssize_t)error;
gfs2_size_hint(inode, *ppos, len);
return generic_file_splice_write(pipe, out, ppos, len, flags);
}
static ssize_t
ncp_file_read(struct file *file, char __user *buf, size_t count, loff_t *ppos)
{
struct dentry *dentry = file->f_path.dentry;
struct inode *inode = dentry->d_inode;
size_t already_read = 0;
off_t pos;
size_t bufsize;
int error;
void* freepage;
size_t freelen;
DPRINTK("ncp_file_read: enter %s/%s\n",
dentry->d_parent->d_name.name, dentry->d_name.name);
pos = *ppos;
if ((ssize_t) count < 0) {
return -EINVAL;
}
if (!count)
return 0;
if (pos > inode->i_sb->s_maxbytes)
return 0;
if (pos + count > inode->i_sb->s_maxbytes) {
count = inode->i_sb->s_maxbytes - pos;
}
error = ncp_make_open(inode, O_RDONLY);
if (error) {
DPRINTK(KERN_ERR "ncp_file_read: open failed, error=%d\n", error);
return error;
}
bufsize = NCP_SERVER(inode)->buffer_size;
error = -EIO;
freelen = ncp_read_bounce_size(bufsize);
freepage = vmalloc(freelen);
if (!freepage)
goto outrel;
error = 0;
/* First read in as much as possible for each bufsize. */
while (already_read < count) {
int read_this_time;
size_t to_read = min_t(unsigned int,
bufsize - (pos % bufsize),
count - already_read);
error = ncp_read_bounce(NCP_SERVER(inode),
NCP_FINFO(inode)->file_handle,
pos, to_read, buf, &read_this_time,
freepage, freelen);
if (error) {
error = -EIO; /* NW errno -> Linux errno */
break;
}
pos += read_this_time;
buf += read_this_time;
already_read += read_this_time;
if (read_this_time != to_read) {
break;
}
}
vfree(freepage);
*ppos = pos;
file_accessed(file);
DPRINTK("ncp_file_read: exit %s/%s\n",
dentry->d_parent->d_name.name, dentry->d_name.name);
outrel:
ncp_inode_close(inode);
return already_read ? already_read : error;
}
static ssize_t
pipe_read(struct kiocb *iocb, const struct iovec *_iov,
unsigned long nr_segs, loff_t pos)
{
struct file *filp = iocb->ki_filp;
struct inode *inode = filp->f_path.dentry->d_inode;
struct pipe_inode_info *pipe;
int do_wakeup;
ssize_t ret;
struct iovec *iov = (struct iovec *)_iov;
size_t total_len;
total_len = iov_length(iov, nr_segs);
/* Null read succeeds. */
if (unlikely(total_len == 0))
return 0;
do_wakeup = 0;
ret = 0;
mutex_lock(&inode->i_mutex);
pipe = inode->i_pipe;
for (;;) {
int bufs = pipe->nrbufs;
if (bufs) {
int curbuf = pipe->curbuf;
struct pipe_buffer *buf = pipe->bufs + curbuf;
const struct pipe_buf_operations *ops = buf->ops;
void *addr;
size_t chars = buf->len;
int error, atomic;
if (chars > total_len)
chars = total_len;
error = ops->confirm(pipe, buf);
if (error) {
if (!ret)
error = ret;
break;
}
atomic = !iov_fault_in_pages_write(iov, chars);
redo:
addr = ops->map(pipe, buf, atomic);
error = pipe_iov_copy_to_user(iov, addr + buf->offset, chars, atomic);
ops->unmap(pipe, buf, addr);
if (unlikely(error)) {
/*
* Just retry with the slow path if we failed.
*/
if (atomic) {
atomic = 0;
goto redo;
}
if (!ret)
ret = error;
break;
}
ret += chars;
buf->offset += chars;
buf->len -= chars;
if (!buf->len) {
buf->ops = NULL;
ops->release(pipe, buf);
curbuf = (curbuf + 1) & (pipe->buffers - 1);
pipe->curbuf = curbuf;
pipe->nrbufs = --bufs;
do_wakeup = 1;
}
total_len -= chars;
if (!total_len)
break; /* common path: read succeeded */
}
if (bufs) /* More to do? */
continue;
if (!pipe->writers)
break;
if (!pipe->waiting_writers) {
/* syscall merging: Usually we must not sleep
* if O_NONBLOCK is set, or if we got some data.
* But if a writer sleeps in kernel space, then
* we can wait for that data without violating POSIX.
*/
if (ret)
break;
if (filp->f_flags & O_NONBLOCK) {
ret = -EAGAIN;
break;
}
}
if (signal_pending(current)) {
if (!ret)
ret = -ERESTARTSYS;
break;
}
if (do_wakeup) {
wake_up_interruptible_sync(&pipe->wait);
kill_fasync(&pipe->fasync_writers, SIGIO, POLL_OUT);
}
pipe_wait(pipe);
}
mutex_unlock(&inode->i_mutex);
/* Signal writers asynchronously that there is more room. */
if (do_wakeup) {
wake_up_interruptible_sync(&pipe->wait);
kill_fasync(&pipe->fasync_writers, SIGIO, POLL_OUT);
}
if (ret > 0)
file_accessed(filp);
return ret;
}
static ssize_t
pipe_readv(struct file *filp, const struct iovec *_iov,
unsigned long nr_segs, loff_t *ppos)
{
struct inode *inode = filp->f_dentry->d_inode;
struct pipe_inode_info *info;
int do_wakeup;
ssize_t ret;
struct iovec *iov = (struct iovec *)_iov;
size_t total_len;
total_len = iov_length(iov, nr_segs);
/* Null read succeeds. */
if (unlikely(total_len == 0))
return 0;
do_wakeup = 0;
ret = 0;
down(PIPE_SEM(*inode));
info = inode->i_pipe;
for (;;) {
int bufs = info->nrbufs;
if (bufs) {
int curbuf = info->curbuf;
struct pipe_buffer *buf = info->bufs + curbuf;
struct pipe_buf_operations *ops = buf->ops;
void *addr;
size_t chars = buf->len;
int error;
if (chars > total_len)
chars = total_len;
addr = ops->map(filp, info, buf);
error = pipe_iov_copy_to_user(iov, addr + buf->offset, chars);
ops->unmap(info, buf);
if (unlikely(error)) {
if (!ret) ret = -EFAULT;
break;
}
ret += chars;
buf->offset += chars;
buf->len -= chars;
if (!buf->len) {
buf->ops = NULL;
ops->release(info, buf);
curbuf = (curbuf + 1) & (PIPE_BUFFERS-1);
info->curbuf = curbuf;
info->nrbufs = --bufs;
do_wakeup = 1;
}
total_len -= chars;
if (!total_len)
break; /* common path: read succeeded */
}
if (bufs) /* More to do? */
continue;
if (!PIPE_WRITERS(*inode))
break;
if (!PIPE_WAITING_WRITERS(*inode)) {
/* syscall merging: Usually we must not sleep
* if O_NONBLOCK is set, or if we got some data.
* But if a writer sleeps in kernel space, then
* we can wait for that data without violating POSIX.
*/
if (ret)
break;
if (filp->f_flags & O_NONBLOCK) {
ret = -EAGAIN;
break;
}
}
if (signal_pending(current)) {
if (!ret) ret = -ERESTARTSYS;
break;
}
if (do_wakeup) {
wake_up_interruptible_sync(PIPE_WAIT(*inode));
kill_fasync(PIPE_FASYNC_WRITERS(*inode), SIGIO, POLL_OUT);
}
pipe_wait(inode);
}
up(PIPE_SEM(*inode));
/* Signal writers asynchronously that there is more room. */
if (do_wakeup) {
wake_up_interruptible(PIPE_WAIT(*inode));
kill_fasync(PIPE_FASYNC_WRITERS(*inode), SIGIO, POLL_OUT);
}
if (ret > 0)
file_accessed(filp);
return ret;
}
