+ errno = pcache_write(inode, buf, count, file->f_pos);
+ }
+
+ if (errno > 0) {
+ file->f_pos += errno;
+ inode->fsize = MAX(inode->fsize, file->f_pos);
+
+ unlock_inode(inode);
+ return errno;
+ }
+
+ unlock_inode(inode);
+
+done:
+ return DO_STATUS(errno);
+}
+
+__DEFINE_LXSYSCALL3(int, lseek, int, fd, int, offset, int, options)
+{
+ int errno = 0;
+ struct v_fd* fd_s;
+ if ((errno = vfs_getfd(fd, &fd_s))) {
+ goto done;
+ }
+
+ struct v_file* file = fd_s->file;
+ struct v_inode* inode = file->inode;
+
+ if (!file->ops->seek) {
+ errno = ENOTSUP;
+ goto done;
+ }
+
+ if (!check_allow_read(inode)) {
+ errno = EPERM;
+ goto done;
+ }
+
+ lock_inode(inode);
+
+ int overflow = 0;
+ int fpos = file->f_pos;
+
+ if (vfs_get_dtype(inode->itype) == DT_DIR) {
+ options = (options != FSEEK_END) ? options : FSEEK_SET;
+ }
+
+ switch (options) {
+ case FSEEK_CUR:
+ overflow = sadd_of((int)file->f_pos, offset, &fpos);
+ break;
+ case FSEEK_END:
+ overflow = sadd_of((int)inode->fsize, offset, &fpos);
+ break;
+ case FSEEK_SET:
+ fpos = offset;
+ break;
+ }
+
+ if (overflow) {
+ errno = EOVERFLOW;
+ }
+ else {
+ errno = file->ops->seek(file, fpos);
+ }
+
+ unlock_inode(inode);
+
+done:
+ return DO_STATUS(errno);
+}
+
+int
+vfs_get_path(struct v_dnode* dnode, char* buf, size_t size, int depth)
+{
+ if (!dnode) {
+ return 0;
+ }
+
+ if (depth > 64) {
+ return ENAMETOOLONG;
+ }
+
+ size_t len = 0;
+
+ if (dnode->parent != dnode) {
+ len = vfs_get_path(dnode->parent, buf, size, depth + 1);
+ }
+
+ if (len >= size) {
+ return len;
+ }
+
+ if (!len || buf[len - 1] != VFS_PATH_DELIM) {
+ buf[len++] = VFS_PATH_DELIM;
+ }
+
+ size_t cpy_size = MIN(dnode->name.len, size - len);
+ strncpy(buf + len, dnode->name.value, cpy_size);
+ len += cpy_size;
+
+ return len;
+}
+
+int
+vfs_readlink(struct v_dnode* dnode, char* buf, size_t size)
+{
+ const char* link;
+ struct v_inode* inode = dnode->inode;
+
+ if (!check_symlink_node(inode)) {
+ return EINVAL;
+ }
+
+ if (!inode->ops->read_symlink) {
+ return ENOTSUP;
+ }
+
+ if (!check_allow_read(inode)) {
+ return EPERM;
+ }
+
+ lock_inode(inode);
+
+ int errno = inode->ops->read_symlink(inode, &link);
+ if (errno >= 0) {
+ strncpy(buf, link, MIN(size, (size_t)errno));
+ }
+
+ unlock_inode(inode);
+ return errno;
+}
+
+int
+vfs_get_dtype(int itype)
+{
+ int dtype = DT_FILE;
+ if (check_itype(itype, VFS_IFSYMLINK)) {
+ dtype |= DT_SYMLINK;
+ }
+
+ if (check_itype(itype, VFS_IFDIR)) {
+ dtype |= DT_DIR;
+ return dtype;
+ }
+
+ // TODO other types
+
+ return dtype;
+}
+
+struct v_fdtable*
+fdtable_create()
+{
+ struct v_fdtable* fdtab;
+
+ fdtab = vzalloc(sizeof(struct v_fdtable));
+ mutex_init(&fdtab->lock);
+
+ return fdtab;
+}
+
+void
+fdtable_copy(struct v_fdtable* dest, struct v_fdtable* src)
+{
+ lock_fdtable(dest);
+ lock_fdtable(src);
+
+ for (size_t i = 0; i < VFS_MAX_FD; i++) {
+ struct v_fd* fd = src->fds[i];
+ if (!fd)
+ continue;
+ vfs_dup_fd(fd, &dest->fds[i]);
+ }
+
+ unlock_fdtable(dest);
+ unlock_fdtable(src);
+}
+
+void
+fdtable_free(struct v_fdtable* table)
+{
+ assert(!mutex_on_hold(&table->lock));
+
+ vfree(table);
+}
+
+__DEFINE_LXSYSCALL3(int, realpathat, int, fd, char*, buf, size_t, size)
+{
+ int errno;
+ struct v_fd* fd_s;
+ if ((errno = vfs_getfd(fd, &fd_s))) {
+ goto done;
+ }
+
+ struct v_dnode* dnode;
+
+ dnode = fd_s->file->dnode;
+
+ lock_dnode(dnode);
+ errno = vfs_get_path(dnode, buf, size, 0);
+ unlock_dnode(dnode);
+
+done:
+ return DO_STATUS(errno);
+}
+
+__DEFINE_LXSYSCALL3(int, readlink, const char*, path, char*, buf, size_t, size)
+{
+ int errno;
+ struct v_dnode* dnode;
+ if (!(errno = vfs_walk_proc(path, &dnode, NULL, VFS_WALK_NOFOLLOW))) {
+ errno = vfs_readlink(dnode, buf, size);
+ }
+
+ if (errno >= 0) {
+ return errno;
+ }
+
+ return DO_STATUS(errno);
+}
+
+__DEFINE_LXSYSCALL4(
+ int, readlinkat, int, dirfd, const char*, pathname, char*, buf, size_t, size)
+{
+ int errno;
+ struct v_fd* fd_s;
+ if ((errno = vfs_getfd(dirfd, &fd_s))) {
+ goto done;
+ }
+
+ pathname = pathname ? pathname : "";
+
+ struct v_dnode* dnode;
+ if (!(errno = vfs_walk(
+ fd_s->file->dnode, pathname, &dnode, NULL, VFS_WALK_NOFOLLOW))) {
+ errno = vfs_readlink(fd_s->file->dnode, buf, size);
+ }
+
+ if (errno >= 0) {
+ return errno;
+ }
+
+done:
+ return DO_STATUS(errno);
+}
+
+/*
+ NOTE
+ When we perform operation that could affect the layout of
+ directory (i.e., rename, mkdir, rmdir). We must lock the parent dir
+ whenever possible. This will blocking any ongoing path walking to reach
+ it hence avoid any partial state.
+*/
+
+__DEFINE_LXSYSCALL1(int, rmdir, const char*, pathname)
+{
+ int errno;
+ struct v_dnode* dnode;
+ if ((errno = vfs_walk_proc(pathname, &dnode, NULL, 0))) {
+ return DO_STATUS(errno);
+ }
+
+ lock_dnode(dnode);
+
+ if (!__check_unlinkable(dnode)) {
+ errno = EPERM;
+ goto done;
+ }
+
+ if ((errno = vfs_check_writable(dnode))) {
+ goto done;
+ }
+
+ if ((dnode->super_block->fs->types & FSTYPE_ROFS)) {
+ errno = EROFS;
+ goto done;
+ }
+
+ if (dnode->ref_count > 1 || dnode->inode->open_count) {
+ errno = EBUSY;
+ goto done;
+ }
+
+ if (!llist_empty(&dnode->children)) {
+ errno = ENOTEMPTY;
+ goto done;
+ }
+
+ struct v_dnode* parent = dnode->parent;
+
+ if (!parent) {
+ errno = EINVAL;
+ goto done;
+ }
+
+ lock_dnode(parent);
+ lock_inode(parent->inode);
+
+ if (check_directory_node(dnode->inode)) {
+ errno = parent->inode->ops->rmdir(parent->inode, dnode);
+ if (!errno) {
+ vfs_dcache_remove(dnode);
+ }
+ } else {
+ errno = ENOTDIR;
+ }
+
+ unlock_inode(parent->inode);
+ unlock_dnode(parent);
+
+done:
+ unlock_dnode(dnode);
+ return DO_STATUS(errno);
+}
+
+__DEFINE_LXSYSCALL1(int, mkdir, const char*, path)
+{
+ int errno;
+ struct hstr name;
+ struct v_inode* inode;
+ struct v_dnode *parent, *dir;
+ char name_value[VFS_NAME_MAXLEN];
+
+ name = HHSTR(name_value, 0, 0);
+
+ if ((errno = vfs_walk_proc(path, &parent, &name, VFS_WALK_PARENT))) {
+ goto done;
+ }
+
+ if (!(errno = vfs_walk(parent, name_value, &dir, NULL, 0))) {
+ errno = EEXIST;
+ goto done;
+ }
+
+ if ((errno = vfs_check_writable(parent))) {
+ goto done;
+ }
+
+ if (!(dir = vfs_d_alloc(parent, &name))) {
+ errno = ENOMEM;
+ goto done;
+ }
+
+ inode = parent->inode;
+
+ lock_dnode(parent);
+ lock_inode(inode);
+
+ if ((parent->super_block->fs->types & FSTYPE_ROFS)) {
+ errno = ENOTSUP;
+ } else if (!inode->ops->mkdir) {
+ errno = ENOTSUP;
+ } else if (!check_directory_node(inode)) {
+ errno = ENOTDIR;
+ } else if (!(errno = inode->ops->mkdir(inode, dir))) {
+ vfs_dcache_add(parent, dir);
+ goto cleanup;
+ }
+
+ vfs_d_free(dir);
+
+cleanup:
+ unlock_inode(inode);
+ unlock_dnode(parent);
+done:
+ return DO_STATUS(errno);
+}
+
+static int
+__vfs_do_unlink(struct v_dnode* dnode)
+{
+ int errno;
+ struct v_inode* inode = dnode->inode;
+
+ if (dnode->ref_count > 1) {
+ return EBUSY;
+ }
+
+ if (!__check_unlinkable(dnode)) {
+ return EPERM;
+ }
+
+ if ((errno = vfs_check_writable(dnode))) {
+ return errno;
+ }
+
+ lock_inode(inode);
+
+ if (inode->open_count) {
+ errno = EBUSY;
+ } else if (!check_directory_node(inode)) {
+ errno = inode->ops->unlink(inode, dnode);
+ if (!errno) {
+ vfs_d_free(dnode);
+ }
+ } else {
+ errno = EISDIR;
+ }
+
+ unlock_inode(inode);
+
+ return errno;
+}
+
+__DEFINE_LXSYSCALL1(int, unlink, const char*, pathname)
+{
+ int errno;
+ struct v_dnode* dnode;
+ if ((errno = vfs_walk_proc(pathname, &dnode, NULL, 0))) {
+ goto done;
+ }
+
+ errno = __vfs_do_unlink(dnode);
+
+done:
+ return DO_STATUS(errno);
+}
+
+__DEFINE_LXSYSCALL2(int, unlinkat, int, fd, const char*, pathname)
+{
+ int errno;
+ struct v_fd* fd_s;
+ if ((errno = vfs_getfd(fd, &fd_s))) {
+ goto done;
+ }
+
+ struct v_dnode* dnode;
+ if (!(errno = vfs_walk(fd_s->file->dnode, pathname, &dnode, NULL, 0))) {
+ errno = __vfs_do_unlink(dnode);
+ }
+
+done:
+ return DO_STATUS(errno);
+}
+
+__DEFINE_LXSYSCALL2(int, link, const char*, oldpath, const char*, newpath)
+{
+ int errno;
+ struct file_locator floc;
+ struct v_dnode *to_link, *name_file;
+
+ errno = __vfs_try_locate_file(oldpath, &floc, 0);
+ if (errno) {
+ goto done;
+ }
+
+ __floc_try_unlock(&floc);
+
+ to_link = floc.file;
+ errno = __vfs_try_locate_file(newpath, &floc, FLOC_MKNAME);
+ if (!errno) {
+ goto done;
+ }
+
+ name_file = floc.file;
+ errno = vfs_link(to_link, name_file);
+ if (errno) {
+ vfs_d_free(name_file);
+ }
+
+done:
+ __floc_try_unlock(&floc);
+ return DO_STATUS(errno);
+}
+
+__DEFINE_LXSYSCALL1(int, fsync, int, fildes)
+{
+ int errno;
+ struct v_fd* fd_s;
+
+ if (!(errno = vfs_getfd(fildes, &fd_s))) {
+ errno = vfs_fsync(fd_s->file);
+ }
+
+ return DO_STATUS(errno);
+}
+
+int
+vfs_dup_fd(struct v_fd* old, struct v_fd** new)
+{
+ int errno = 0;
+ struct v_fd* copied = cake_grab(fd_pile);
+
+ memcpy(copied, old, sizeof(struct v_fd));
+
+ vfs_ref_file(old->file);
+
+ *new = copied;
+
+ return errno;
+}
+
+int
+vfs_dup2(int oldfd, int newfd)
+{
+ int errno;
+ struct v_fdtable* fdtab;
+ struct v_fd *oldfd_s, *newfd_s;
+
+ if (newfd == oldfd) {
+ return newfd;
+ }
+
+ if ((errno = vfs_getfd(oldfd, &oldfd_s))) {
+ goto done;
+ }
+
+ if (!TEST_FD(newfd)) {
+ errno = EBADF;
+ goto done;
+ }
+
+ fdtab = __current->fdtable;
+ lock_fdtable(fdtab);
+
+ newfd_s = fdtab->fds[newfd];
+ if (newfd_s && (errno = vfs_close(newfd_s->file))) {
+ goto unlock_and_done;
+ }
+
+ if ((errno = vfs_dup_fd(oldfd_s, &newfd_s))) {
+ goto unlock_and_done;
+ }
+
+ fdtab->fds[newfd] = newfd_s;
+
+ unlock_fdtable(fdtab);
+ return newfd;
+
+unlock_and_done:
+ unlock_fdtable(fdtab);
+
+done:
+ return DO_STATUS(errno);
+}
+
+__DEFINE_LXSYSCALL2(int, dup2, int, oldfd, int, newfd)
+{
+ return vfs_dup2(oldfd, newfd);
+}
+
+__DEFINE_LXSYSCALL1(int, dup, int, oldfd)
+{
+ int errno, newfd;
+ struct v_fd *oldfd_s, *newfd_s;
+ if ((errno = vfs_getfd(oldfd, &oldfd_s))) {
+ goto done;
+ }
+
+ if (!(errno = vfs_alloc_fdslot(&newfd)) &&
+ !(errno = vfs_dup_fd(oldfd_s, &newfd_s))) {
+ __current->fdtable->fds[newfd] = newfd_s;
+ return newfd;