static int gfs2_page_mkwrite(struct vm_fault *vmf) { struct page *page = vmf->page; struct inode *inode = file_inode(vmf->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_SHIFT; unsigned int data_blocks, ind_blocks, rblocks; struct gfs2_holder gh; loff_t size; int ret; sb_start_pagefault(inode->i_sb); ret = gfs2_rsqa_alloc(ip); if (ret) goto out; gfs2_size_hint(vmf->vma->vm_file, pos, PAGE_SIZE); gfs2_holder_init(ip->i_gl, LM_ST_EXCLUSIVE, 0, &gh); ret = gfs2_glock_nq(&gh); if (ret) goto out_uninit; /* Update file times before taking page lock */ file_update_time(vmf->vma->vm_file); 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_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; gfs2_write_calc_reserv(ip, PAGE_SIZE, &data_blocks, &ind_blocks); ap.target = data_blocks + ind_blocks; ret = gfs2_quota_lock_check(ip, &ap); if (ret) goto out_unlock; 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_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: 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, }; /** * 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_rsqa_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_ALL; 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 | I_DIRTY_TIME); if (sync_state) { ret = sync_inode_metadata(inode, 1); if (ret) return ret; if (gfs2_is_jdata(ip)) ret = file_write_and_wait(file); if (ret) return ret; gfs2_ail_flush(ip->i_gl, 1); } if (mapping->nrpages) ret = file_fdatawait_range(file, start, end); return ret ? ret : ret1; } /** * gfs2_file_write_iter - 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_write_iter(struct kiocb *iocb, struct iov_iter *from) { struct file *file = iocb->ki_filp; struct gfs2_inode *ip = GFS2_I(file_inode(file)); int ret; ret = gfs2_rsqa_alloc(ip); if (ret) return ret; gfs2_size_hint(file, iocb->ki_pos, iov_iter_count(from)); if (iocb->ki_flags & IOCB_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_write_iter(iocb, from); } 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; 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; } } out: brelse(dibh); return error; } /** * calc_max_reserv() - Reverse of write_calc_reserv. Given a number of * blocks, determine how many bytes can be written. * @ip: The inode in question. * @len: Max cap of bytes. What we return in *len must be <= this. * @data_blocks: Compute and return the number of data blocks needed * @ind_blocks: Compute and return the number of indirect blocks needed * @max_blocks: The total blocks available to work with. * * Returns: void, but @len, @data_blocks and @ind_blocks are filled in. */ static void calc_max_reserv(struct gfs2_inode *ip, loff_t *len, unsigned int *data_blocks, unsigned int *ind_blocks, unsigned int max_blocks) { loff_t max = *len; const struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode); 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; } *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, max_blks = UINT_MAX; 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; next = (next + 1) << sdp->sd_sb.sb_bsize_shift; 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; gfs2_size_hint(file, offset, len); gfs2_write_calc_reserv(ip, PAGE_SIZE, &data_blocks, &ind_blocks); ap.min_target = data_blocks + ind_blocks; while (len > 0) { if (len < bytes) bytes = len; if (!gfs2_write_alloc_required(ip, offset, bytes)) { len -= bytes; offset += bytes; continue; } /* We need to determine how many bytes we can actually * fallocate without exceeding quota or going over the * end of the fs. We start off optimistically by assuming * we can write max_bytes */ max_bytes = (len > max_chunk_size) ? max_chunk_size : len; /* Since max_bytes is most likely a theoretical max, we * calculate a more realistic 'bytes' to serve as a good * starting point for the number of bytes we may be able * to write */ gfs2_write_calc_reserv(ip, bytes, &data_blocks, &ind_blocks); ap.target = data_blocks + ind_blocks; error = gfs2_quota_lock_check(ip, &ap); if (error) return error; /* ap.allowed tells us how many blocks quota will allow * us to write. Check if this reduces max_blks */ if (ap.allowed && ap.allowed < max_blks) max_blks = ap.allowed; error = gfs2_inplace_reserve(ip, &ap); if (error) goto out_qunlock; /* check if the selected rgrp limits our max_blks further */ if (ap.allowed && ap.allowed < max_blks) max_blks = ap.allowed; /* Almost done. Calculate bytes that can be written using * max_blks. We also recompute max_bytes, data_blocks and * ind_blocks */ calc_max_reserv(ip, &max_bytes, &data_blocks, &ind_blocks, max_blks); 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_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 (!(mode & FALLOC_FL_KEEP_SIZE) && (pos + count) > inode->i_size) { i_size_write(inode, pos + count); file_update_time(file); mark_inode_dirty(inode); } if ((file->f_flags & O_DSYNC) || IS_SYNC(file->f_mapping->host)) return vfs_fsync_range(file, pos, pos + count - 1, (file->f_flags & __O_SYNC) ? 0 : 1); return 0; out_trans_fail: gfs2_inplace_release(ip); out_qunlock: gfs2_quota_unlock(ip); return error; } 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_holder gh; int ret; if (mode & ~FALLOC_FL_KEEP_SIZE) return -EOPNOTSUPP; /* fallocate is needed by gfs2_grow to reserve space in the rindex */ if (gfs2_is_jdata(ip) && inode != sdp->sd_rindex) return -EOPNOTSUPP; inode_lock(inode); gfs2_holder_init(ip->i_gl, LM_ST_EXCLUSIVE, 0, &gh); ret = gfs2_glock_nq(&gh); if (ret) goto out_uninit; if (!(mode & FALLOC_FL_KEEP_SIZE) && (offset + len) > inode->i_size) { ret = inode_newsize_ok(inode, offset + len); if (ret) goto out_unlock; } ret = get_write_access(inode); if (ret) goto out_unlock; ret = gfs2_rsqa_alloc(ip); if (ret) goto out_putw; ret = __gfs2_fallocate(file, mode, offset, len); if (ret) gfs2_rs_deltree(&ip->i_res); out_putw: put_write_access(inode); out_unlock: gfs2_glock_dq(&gh); out_uninit: gfs2_holder_uninit(&gh); inode_unlock(inode); return ret; } 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 gfs2_inode *ip = GFS2_I(out->f_mapping->host); error = gfs2_rsqa_alloc(ip); if (error) return (ssize_t)error; gfs2_size_hint(out, *ppos, len); return iter_file_splice_write(pipe, out, ppos, len, flags); } #ifdef CONFIG_GFS2_FS_LOCKING_DLM /** * 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) locks_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; u16 flags; int error = 0; int sleeptime; state = (fl->fl_type == F_WRLCK) ? LM_ST_EXCLUSIVE : LM_ST_SHARED; flags = (IS_SETLKW(cmd) ? 0 : LM_FLAG_TRY_1CB) | GL_EXACT; mutex_lock(&fp->f_fl_mutex); if (gfs2_holder_initialized(fl_gh)) { if (fl_gh->gh_state == state) goto out; locks_lock_file_wait(file, &(struct file_lock) { .fl_type = F_UNLCK, .fl_flags = FL_FLOCK }); gfs2_glock_dq(fl_gh); gfs2_holder_reinit(state, flags, fl_gh); } else {
static int link_dinode(struct gfs2_inode *dip, const struct qstr *name, struct gfs2_inode *ip, struct gfs2_diradd *da) { struct gfs2_sbd *sdp = GFS2_SB(&dip->i_inode); struct gfs2_alloc_parms ap = { .target = da->nr_blocks, }; int error; if (da->nr_blocks) { error = gfs2_quota_lock_check(dip, &ap); if (error) goto fail_quota_locks; error = gfs2_inplace_reserve(dip, &ap); if (error) goto fail_quota_locks; error = gfs2_trans_begin(sdp, gfs2_trans_da_blks(dip, da, 2), 0); if (error) goto fail_ipreserv; } else { error = gfs2_trans_begin(sdp, RES_LEAF + 2 * RES_DINODE, 0); if (error) goto fail_quota_locks; } error = gfs2_dir_add(&dip->i_inode, name, ip, da); gfs2_trans_end(sdp); fail_ipreserv: gfs2_inplace_release(dip); fail_quota_locks: gfs2_quota_unlock(dip); return error; } static int gfs2_initxattrs(struct inode *inode, const struct xattr *xattr_array, void *fs_info) { const struct xattr *xattr; int err = 0; for (xattr = xattr_array; xattr->name != NULL; xattr++) { err = __gfs2_xattr_set(inode, xattr->name, xattr->value, xattr->value_len, 0, GFS2_EATYPE_SECURITY); if (err < 0) break; } return err; } /** * gfs2_create_inode - Create a new inode * @dir: The parent directory * @dentry: The new dentry * @file: If non-NULL, the file which is being opened * @mode: The permissions on the new inode * @dev: For device nodes, this is the device number * @symname: For symlinks, this is the link destination * @size: The initial size of the inode (ignored for directories) * * Returns: 0 on success, or error code */ static int gfs2_create_inode(struct inode *dir, struct dentry *dentry, struct file *file, umode_t mode, dev_t dev, const char *symname, unsigned int size, int excl) { const struct qstr *name = &dentry->d_name; struct posix_acl *default_acl, *acl; struct gfs2_holder ghs[2]; struct inode *inode = NULL; struct gfs2_inode *dip = GFS2_I(dir), *ip; struct gfs2_sbd *sdp = GFS2_SB(&dip->i_inode); struct gfs2_glock *io_gl = NULL; int error, free_vfs_inode = 1; u32 aflags = 0; unsigned blocks = 1; struct gfs2_diradd da = { .bh = NULL, .save_loc = 1, }; if (!name->len || name->len > GFS2_FNAMESIZE) return -ENAMETOOLONG; error = gfs2_rsqa_alloc(dip); if (error) return error; error = gfs2_rindex_update(sdp); if (error) return error; error = gfs2_glock_nq_init(dip->i_gl, LM_ST_EXCLUSIVE, 0, ghs); if (error) goto fail; gfs2_holder_mark_uninitialized(ghs + 1); error = create_ok(dip, name, mode); if (error) goto fail_gunlock; inode = gfs2_dir_search(dir, &dentry->d_name, !S_ISREG(mode) || excl); error = PTR_ERR(inode); if (!IS_ERR(inode)) { if (S_ISDIR(inode->i_mode)) { iput(inode); inode = ERR_PTR(-EISDIR); goto fail_gunlock; } d_instantiate(dentry, inode); error = 0; if (file) { if (S_ISREG(inode->i_mode)) error = finish_open(file, dentry, gfs2_open_common); else error = finish_no_open(file, NULL); } gfs2_glock_dq_uninit(ghs); return error; } else if (error != -ENOENT) { goto fail_gunlock; } error = gfs2_diradd_alloc_required(dir, name, &da); if (error < 0) goto fail_gunlock; inode = new_inode(sdp->sd_vfs); error = -ENOMEM; if (!inode) goto fail_gunlock; error = posix_acl_create(dir, &mode, &default_acl, &acl); if (error) goto fail_gunlock; ip = GFS2_I(inode); error = gfs2_rsqa_alloc(ip); if (error) goto fail_free_acls; inode->i_mode = mode; set_nlink(inode, S_ISDIR(mode) ? 2 : 1); inode->i_rdev = dev; inode->i_size = size; inode->i_atime = inode->i_mtime = inode->i_ctime = current_time(inode); gfs2_set_inode_blocks(inode, 1); munge_mode_uid_gid(dip, inode); check_and_update_goal(dip); ip->i_goal = dip->i_goal; ip->i_diskflags = 0; ip->i_eattr = 0; ip->i_height = 0; ip->i_depth = 0; ip->i_entries = 0; ip->i_no_addr = 0; /* Temporarily zero until real addr is assigned */ switch(mode & S_IFMT) { case S_IFREG: if ((dip->i_diskflags & GFS2_DIF_INHERIT_JDATA) || gfs2_tune_get(sdp, gt_new_files_jdata)) ip->i_diskflags |= GFS2_DIF_JDATA; gfs2_set_aops(inode); break; case S_IFDIR: ip->i_diskflags |= (dip->i_diskflags & GFS2_DIF_INHERIT_JDATA); ip->i_diskflags |= GFS2_DIF_JDATA; ip->i_entries = 2; break; } /* Force SYSTEM flag on all files and subdirs of a SYSTEM directory */ if (dip->i_diskflags & GFS2_DIF_SYSTEM) ip->i_diskflags |= GFS2_DIF_SYSTEM; gfs2_set_inode_flags(inode); if ((GFS2_I(d_inode(sdp->sd_root_dir)) == dip) || (dip->i_diskflags & GFS2_DIF_TOPDIR)) aflags |= GFS2_AF_ORLOV; if (default_acl || acl) blocks++; error = alloc_dinode(ip, aflags, &blocks); if (error) goto fail_free_inode; gfs2_set_inode_blocks(inode, blocks); error = gfs2_glock_get(sdp, ip->i_no_addr, &gfs2_inode_glops, CREATE, &ip->i_gl); if (error) goto fail_free_inode; flush_delayed_work(&ip->i_gl->gl_work); glock_set_object(ip->i_gl, ip); error = gfs2_glock_nq_init(ip->i_gl, LM_ST_EXCLUSIVE, GL_SKIP, ghs + 1); if (error) goto fail_free_inode; error = gfs2_trans_begin(sdp, blocks, 0); if (error) goto fail_gunlock2; if (blocks > 1) { ip->i_eattr = ip->i_no_addr + 1; gfs2_init_xattr(ip); } init_dinode(dip, ip, symname); gfs2_trans_end(sdp); error = gfs2_glock_get(sdp, ip->i_no_addr, &gfs2_iopen_glops, CREATE, &io_gl); if (error) goto fail_gunlock2; BUG_ON(test_and_set_bit(GLF_INODE_CREATING, &io_gl->gl_flags)); error = gfs2_glock_nq_init(io_gl, LM_ST_SHARED, GL_EXACT, &ip->i_iopen_gh); if (error) goto fail_gunlock2; glock_set_object(ip->i_iopen_gh.gh_gl, ip); gfs2_glock_put(io_gl); gfs2_set_iop(inode); insert_inode_hash(inode); free_vfs_inode = 0; /* After this point, the inode is no longer considered free. Any failures need to undo the gfs2 structures. */ if (default_acl) { error = __gfs2_set_acl(inode, default_acl, ACL_TYPE_DEFAULT); if (error) goto fail_gunlock3; posix_acl_release(default_acl); default_acl = NULL; } if (acl) { error = __gfs2_set_acl(inode, acl, ACL_TYPE_ACCESS); if (error) goto fail_gunlock3; posix_acl_release(acl); acl = NULL; } error = security_inode_init_security(&ip->i_inode, &dip->i_inode, name, &gfs2_initxattrs, NULL); if (error) goto fail_gunlock3; error = link_dinode(dip, name, ip, &da); if (error) goto fail_gunlock3; mark_inode_dirty(inode); d_instantiate(dentry, inode); if (file) { file->f_mode |= FMODE_CREATED; error = finish_open(file, dentry, gfs2_open_common); } gfs2_glock_dq_uninit(ghs); gfs2_glock_dq_uninit(ghs + 1); clear_bit(GLF_INODE_CREATING, &io_gl->gl_flags); return error; fail_gunlock3: glock_clear_object(io_gl, ip); gfs2_glock_dq_uninit(&ip->i_iopen_gh); gfs2_glock_put(io_gl); fail_gunlock2: if (io_gl) clear_bit(GLF_INODE_CREATING, &io_gl->gl_flags); fail_free_inode: if (ip->i_gl) { glock_clear_object(ip->i_gl, ip); gfs2_glock_put(ip->i_gl); } gfs2_rsqa_delete(ip, NULL); fail_free_acls: posix_acl_release(default_acl); posix_acl_release(acl); fail_gunlock: gfs2_dir_no_add(&da); gfs2_glock_dq_uninit(ghs); if (inode && !IS_ERR(inode)) { clear_nlink(inode); if (!free_vfs_inode) mark_inode_dirty(inode); set_bit(free_vfs_inode ? GIF_FREE_VFS_INODE : GIF_ALLOC_FAILED, &GFS2_I(inode)->i_flags); iput(inode); } if (gfs2_holder_initialized(ghs + 1)) gfs2_glock_dq_uninit(ghs + 1); fail: return error; } /** * gfs2_create - Create a file * @dir: The directory in which to create the file * @dentry: The dentry of the new file * @mode: The mode of the new file * * Returns: errno */ static int gfs2_create(struct inode *dir, struct dentry *dentry, umode_t mode, bool excl) { return gfs2_create_inode(dir, dentry, NULL, S_IFREG | mode, 0, NULL, 0, excl); } /** * __gfs2_lookup - Look up a filename in a directory and return its inode * @dir: The directory inode * @dentry: The dentry of the new inode * @file: File to be opened * * * Returns: errno */ static struct dentry *__gfs2_lookup(struct inode *dir, struct dentry *dentry, struct file *file) { struct inode *inode; struct dentry *d; struct gfs2_holder gh; struct gfs2_glock *gl; int error; inode = gfs2_lookupi(dir, &dentry->d_name, 0); if (inode == NULL) { d_add(dentry, NULL); return NULL; } if (IS_ERR(inode)) return ERR_CAST(inode); gl = GFS2_I(inode)->i_gl; error = gfs2_glock_nq_init(gl, LM_ST_SHARED, LM_FLAG_ANY, &gh); if (error) { iput(inode); return ERR_PTR(error); } d = d_splice_alias(inode, dentry); if (IS_ERR(d)) { gfs2_glock_dq_uninit(&gh); return d; } if (file && S_ISREG(inode->i_mode)) error = finish_open(file, dentry, gfs2_open_common); gfs2_glock_dq_uninit(&gh); if (error) { dput(d); return ERR_PTR(error); } return d; } static struct dentry *gfs2_lookup(struct inode *dir, struct dentry *dentry, unsigned flags) { return __gfs2_lookup(dir, dentry, NULL); } /** * gfs2_link - Link to a file * @old_dentry: The inode to link * @dir: Add link to this directory * @dentry: The name of the link * * Link the inode in "old_dentry" into the directory "dir" with the * name in "dentry". * * Returns: errno */ static int gfs2_link(struct dentry *old_dentry, struct inode *dir, struct dentry *dentry) { struct gfs2_inode *dip = GFS2_I(dir); struct gfs2_sbd *sdp = GFS2_SB(dir); struct inode *inode = d_inode(old_dentry); struct gfs2_inode *ip = GFS2_I(inode); struct gfs2_holder ghs[2]; struct buffer_head *dibh; struct gfs2_diradd da = { .bh = NULL, .save_loc = 1, }; int error; if (S_ISDIR(inode->i_mode)) return -EPERM; error = gfs2_rsqa_alloc(dip); if (error) return error; gfs2_holder_init(dip->i_gl, LM_ST_EXCLUSIVE, 0, ghs); gfs2_holder_init(ip->i_gl, LM_ST_EXCLUSIVE, 0, ghs + 1); error = gfs2_glock_nq(ghs); /* parent */ if (error) goto out_parent; error = gfs2_glock_nq(ghs + 1); /* child */ if (error) goto out_child; error = -ENOENT; if (inode->i_nlink == 0) goto out_gunlock; error = gfs2_permission(dir, MAY_WRITE | MAY_EXEC); if (error) goto out_gunlock; error = gfs2_dir_check(dir, &dentry->d_name, NULL); switch (error) { case -ENOENT: break; case 0: error = -EEXIST; default: goto out_gunlock; } error = -EINVAL; if (!dip->i_inode.i_nlink) goto out_gunlock; error = -EFBIG; if (dip->i_entries == (u32)-1) goto out_gunlock; error = -EPERM; if (IS_IMMUTABLE(inode) || IS_APPEND(inode)) goto out_gunlock; error = -EINVAL; if (!ip->i_inode.i_nlink) goto out_gunlock; error = -EMLINK; if (ip->i_inode.i_nlink == (u32)-1) goto out_gunlock; error = gfs2_diradd_alloc_required(dir, &dentry->d_name, &da); if (error < 0) goto out_gunlock; if (da.nr_blocks) { struct gfs2_alloc_parms ap = { .target = da.nr_blocks, }; error = gfs2_quota_lock_check(dip, &ap); if (error) goto out_gunlock; error = gfs2_inplace_reserve(dip, &ap); if (error) goto out_gunlock_q; error = gfs2_trans_begin(sdp, gfs2_trans_da_blks(dip, &da, 2), 0); if (error) goto out_ipres; } else { error = gfs2_trans_begin(sdp, 2 * RES_DINODE + RES_LEAF, 0); if (error) goto out_ipres; } error = gfs2_meta_inode_buffer(ip, &dibh); if (error) goto out_end_trans; error = gfs2_dir_add(dir, &dentry->d_name, ip, &da); if (error) goto out_brelse; gfs2_trans_add_meta(ip->i_gl, dibh); inc_nlink(&ip->i_inode); ip->i_inode.i_ctime = current_time(&ip->i_inode); ihold(inode); d_instantiate(dentry, inode); mark_inode_dirty(inode); out_brelse: brelse(dibh); out_end_trans: gfs2_trans_end(sdp); out_ipres: if (da.nr_blocks) gfs2_inplace_release(dip); out_gunlock_q: if (da.nr_blocks) gfs2_quota_unlock(dip); out_gunlock: gfs2_dir_no_add(&da); gfs2_glock_dq(ghs + 1); out_child: gfs2_glock_dq(ghs); out_parent: gfs2_holder_uninit(ghs); gfs2_holder_uninit(ghs + 1); return error; } /* * gfs2_unlink_ok - check to see that a inode is still in a directory * @dip: the directory * @name: the name of the file * @ip: the inode * * Assumes that the lock on (at least) @dip is held. * * Returns: 0 if the parent/child relationship is correct, errno if it isn't */ static int gfs2_unlink_ok(struct gfs2_inode *dip, const struct qstr *name, const struct gfs2_inode *ip) { int error; if (IS_IMMUTABLE(&ip->i_inode) || IS_APPEND(&ip->i_inode)) return -EPERM; if ((dip->i_inode.i_mode & S_ISVTX) && !uid_eq(dip->i_inode.i_uid, current_fsuid()) && !uid_eq(ip->i_inode.i_uid, current_fsuid()) && !capable(CAP_FOWNER)) return -EPERM; if (IS_APPEND(&dip->i_inode)) return -EPERM; error = gfs2_permission(&dip->i_inode, MAY_WRITE | MAY_EXEC); if (error) return error; return gfs2_dir_check(&dip->i_inode, name, ip); } /** * gfs2_unlink_inode - Removes an inode from its parent dir and unlinks it * @dip: The parent directory * @name: The name of the entry in the parent directory * @inode: The inode to be removed * * Called with all the locks and in a transaction. This will only be * called for a directory after it has been checked to ensure it is empty. * * Returns: 0 on success, or an error */ static int gfs2_unlink_inode(struct gfs2_inode *dip, const struct dentry *dentry) { struct inode *inode = d_inode(dentry); struct gfs2_inode *ip = GFS2_I(inode); int error; error = gfs2_dir_del(dip, dentry); if (error) return error; ip->i_entries = 0; inode->i_ctime = current_time(inode); if (S_ISDIR(inode->i_mode)) clear_nlink(inode); else drop_nlink(inode); mark_inode_dirty(inode); if (inode->i_nlink == 0) gfs2_unlink_di(inode); return 0; } /** * gfs2_unlink - Unlink an inode (this does rmdir as well) * @dir: The inode of the directory containing the inode to unlink * @dentry: The file itself * * This routine uses the type of the inode as a flag to figure out * whether this is an unlink or an rmdir. * * Returns: errno */ static int gfs2_unlink(struct inode *dir, struct dentry *dentry) { struct gfs2_inode *dip = GFS2_I(dir); struct gfs2_sbd *sdp = GFS2_SB(dir); struct inode *inode = d_inode(dentry); struct gfs2_inode *ip = GFS2_I(inode); struct gfs2_holder ghs[3]; struct gfs2_rgrpd *rgd; int error; error = gfs2_rindex_update(sdp); if (error) return error; error = -EROFS; gfs2_holder_init(dip->i_gl, LM_ST_EXCLUSIVE, 0, ghs); gfs2_holder_init(ip->i_gl, LM_ST_EXCLUSIVE, 0, ghs + 1); rgd = gfs2_blk2rgrpd(sdp, ip->i_no_addr, 1); if (!rgd) goto out_inodes; gfs2_holder_init(rgd->rd_gl, LM_ST_EXCLUSIVE, 0, ghs + 2); error = gfs2_glock_nq(ghs); /* parent */ if (error) goto out_parent; error = gfs2_glock_nq(ghs + 1); /* child */ if (error) goto out_child; error = -ENOENT; if (inode->i_nlink == 0) goto out_rgrp; if (S_ISDIR(inode->i_mode)) { error = -ENOTEMPTY; if (ip->i_entries > 2 || inode->i_nlink > 2) goto out_rgrp; } error = gfs2_glock_nq(ghs + 2); /* rgrp */ if (error) goto out_rgrp; error = gfs2_unlink_ok(dip, &dentry->d_name, ip); if (error) goto out_gunlock; error = gfs2_trans_begin(sdp, 2*RES_DINODE + 3*RES_LEAF + RES_RG_BIT, 0); if (error) goto out_gunlock; error = gfs2_unlink_inode(dip, dentry); gfs2_trans_end(sdp); out_gunlock: gfs2_glock_dq(ghs + 2); out_rgrp: gfs2_glock_dq(ghs + 1); out_child: gfs2_glock_dq(ghs); out_parent: gfs2_holder_uninit(ghs + 2); out_inodes: gfs2_holder_uninit(ghs + 1); gfs2_holder_uninit(ghs); return error; } /** * gfs2_symlink - Create a symlink * @dir: The directory to create the symlink in * @dentry: The dentry to put the symlink in * @symname: The thing which the link points to * * Returns: errno */ static int gfs2_symlink(struct inode *dir, struct dentry *dentry, const char *symname) { unsigned int size; size = strlen(symname); if (size >= gfs2_max_stuffed_size(GFS2_I(dir))) return -ENAMETOOLONG; return gfs2_create_inode(dir, dentry, NULL, S_IFLNK | S_IRWXUGO, 0, symname, size, 0); } /** * gfs2_mkdir - Make a directory * @dir: The parent directory of the new one * @dentry: The dentry of the new directory * @mode: The mode of the new directory * * Returns: errno */ static int gfs2_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode) { unsigned dsize = gfs2_max_stuffed_size(GFS2_I(dir)); return gfs2_create_inode(dir, dentry, NULL, S_IFDIR | mode, 0, NULL, dsize, 0); }