Exemple #1
0
/**
 * generic_file_splice_write_nolock - generic_file_splice_write without mutexes
 * @pipe:	pipe info
 * @out:	file to write to
 * @len:	number of bytes to splice
 * @flags:	splice modifier flags
 *
 * Will either move or copy pages (determined by @flags options) from
 * the given pipe inode to the given file. The caller is responsible
 * for acquiring i_mutex on both inodes.
 *
 */
ssize_t
generic_file_splice_write_nolock(struct pipe_inode_info *pipe, struct file *out,
				 loff_t *ppos, size_t len, unsigned int flags)
{
	struct address_space *mapping = out->f_mapping;
	struct inode *inode = mapping->host;
	ssize_t ret;
	int err;

	err = remove_suid(out->f_dentry);
	if (unlikely(err))
		return err;

	ret = __splice_from_pipe(pipe, out, ppos, len, flags, pipe_to_file);
	if (ret > 0) {
		*ppos += ret;

		/*
		 * If file or inode is SYNC and we actually wrote some data,
		 * sync it.
		 */
		if (unlikely((out->f_flags & O_SYNC) || IS_SYNC(inode))) {
			err = generic_osync_inode(inode, mapping,
						  OSYNC_METADATA|OSYNC_DATA);

			if (err)
				ret = err;
		}
	}

	return ret;
}
Exemple #2
0
/**
 * generic_file_splice_write_nolock - generic_file_splice_write without mutexes
 * @pipe:	pipe info
 * @out:	file to write to
 * @len:	number of bytes to splice
 * @flags:	splice modifier flags
 *
 * Will either move or copy pages (determined by @flags options) from
 * the given pipe inode to the given file. The caller is responsible
 * for acquiring i_mutex on both inodes.
 *
 */
ssize_t
generic_file_splice_write_nolock(struct pipe_inode_info *pipe, struct file *out,
				 loff_t *ppos, size_t len, unsigned int flags)
{
	struct address_space *mapping = out->f_mapping;
	struct inode *inode = mapping->host;
	ssize_t ret;
	int err;

	err = remove_suid(out->f_path.dentry);
	if (unlikely(err))
		return err;

	ret = __splice_from_pipe(pipe, out, ppos, len, flags, pipe_to_file);
	if (ret > 0) {
		unsigned long nr_pages;

		*ppos += ret;
		nr_pages = (ret + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;

		/*
		 * If file or inode is SYNC and we actually wrote some data,
		 * sync it.
		 */
		if (unlikely((out->f_flags & O_SYNC) || IS_SYNC(inode))) {
			err = generic_osync_inode(inode, mapping,
						  OSYNC_METADATA|OSYNC_DATA);

			if (err)
				ret = err;
		}
		balance_dirty_pages_ratelimited_nr(mapping, nr_pages);
	}
Exemple #3
0
ssize_t				/* bytes written, or (-) error */
xfs_write(
	bhv_desc_t		*bdp,
	struct kiocb		*iocb,
	const struct iovec	*iovp,
	unsigned int		nsegs,
	loff_t			*offset,
	int			ioflags,
	cred_t			*credp)
{
	struct file		*file = iocb->ki_filp;
	struct address_space	*mapping = file->f_mapping;
	struct inode		*inode = mapping->host;
	unsigned long		segs = nsegs;
	xfs_inode_t		*xip;
	xfs_mount_t		*mp;
	ssize_t			ret = 0, error = 0;
	xfs_fsize_t		isize, new_size;
	xfs_iocore_t		*io;
	vnode_t			*vp;
	unsigned long		seg;
	int			iolock;
	int			eventsent = 0;
	vrwlock_t		locktype;
	size_t			ocount = 0, count;
	loff_t			pos;
	int			need_isem = 1, need_flush = 0;

	XFS_STATS_INC(xs_write_calls);

	vp = BHV_TO_VNODE(bdp);
	xip = XFS_BHVTOI(bdp);

	for (seg = 0; seg < segs; seg++) {
		const struct iovec *iv = &iovp[seg];

		/*
		 * If any segment has a negative length, or the cumulative
		 * length ever wraps negative then return -EINVAL.
		 */
		ocount += iv->iov_len;
		if (unlikely((ssize_t)(ocount|iv->iov_len) < 0))
			return -EINVAL;
		if (access_ok(VERIFY_READ, iv->iov_base, iv->iov_len))
			continue;
		if (seg == 0)
			return -EFAULT;
		segs = seg;
		ocount -= iv->iov_len;  /* This segment is no good */
		break;
	}

	count = ocount;
	pos = *offset;

	if (count == 0)
		return 0;

	io = &xip->i_iocore;
	mp = io->io_mount;

	if (XFS_FORCED_SHUTDOWN(mp))
		return -EIO;

	if (ioflags & IO_ISDIRECT) {
		xfs_buftarg_t	*target =
			(xip->i_d.di_flags & XFS_DIFLAG_REALTIME) ?
				mp->m_rtdev_targp : mp->m_ddev_targp;

		if ((pos & target->pbr_smask) || (count & target->pbr_smask))
			return XFS_ERROR(-EINVAL);

		if (!VN_CACHED(vp) && pos < i_size_read(inode))
			need_isem = 0;

		if (VN_CACHED(vp))
			need_flush = 1;
	}

relock:
	if (need_isem) {
		iolock = XFS_IOLOCK_EXCL;
		locktype = VRWLOCK_WRITE;

		down(&inode->i_sem);
	} else {
		iolock = XFS_IOLOCK_SHARED;
		locktype = VRWLOCK_WRITE_DIRECT;
	}

	xfs_ilock(xip, XFS_ILOCK_EXCL|iolock);

	isize = i_size_read(inode);

	if (file->f_flags & O_APPEND)
		*offset = isize;

start:
	error = -generic_write_checks(file, &pos, &count,
					S_ISBLK(inode->i_mode));
	if (error) {
		xfs_iunlock(xip, XFS_ILOCK_EXCL|iolock);
		goto out_unlock_isem;
	}

	new_size = pos + count;
	if (new_size > isize)
		io->io_new_size = new_size;

	if ((DM_EVENT_ENABLED(vp->v_vfsp, xip, DM_EVENT_WRITE) &&
	    !(ioflags & IO_INVIS) && !eventsent)) {
		loff_t		savedsize = pos;
		int		dmflags = FILP_DELAY_FLAG(file);

		if (need_isem)
			dmflags |= DM_FLAGS_ISEM;

		xfs_iunlock(xip, XFS_ILOCK_EXCL);
		error = XFS_SEND_DATA(xip->i_mount, DM_EVENT_WRITE, vp,
				      pos, count,
				      dmflags, &locktype);
		if (error) {
			xfs_iunlock(xip, iolock);
			goto out_unlock_isem;
		}
		xfs_ilock(xip, XFS_ILOCK_EXCL);
		eventsent = 1;

		/*
		 * The iolock was dropped and reaquired in XFS_SEND_DATA
		 * so we have to recheck the size when appending.
		 * We will only "goto start;" once, since having sent the
		 * event prevents another call to XFS_SEND_DATA, which is
		 * what allows the size to change in the first place.
		 */
		if ((file->f_flags & O_APPEND) && savedsize != isize) {
			pos = isize = xip->i_d.di_size;
			goto start;
		}
	}

	/*
	 * On Linux, generic_file_write updates the times even if
	 * no data is copied in so long as the write had a size.
	 *
	 * We must update xfs' times since revalidate will overcopy xfs.
	 */
	if (!(ioflags & IO_INVIS)) {
		xfs_ichgtime(xip, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
		inode_update_time(inode, 1);
	}

	/*
	 * If the offset is beyond the size of the file, we have a couple
	 * of things to do. First, if there is already space allocated
	 * we need to either create holes or zero the disk or ...
	 *
	 * If there is a page where the previous size lands, we need
	 * to zero it out up to the new size.
	 */

	if (pos > isize) {
		error = xfs_zero_eof(BHV_TO_VNODE(bdp), io, pos,
					isize, pos + count);
		if (error) {
			xfs_iunlock(xip, XFS_ILOCK_EXCL|iolock);
			goto out_unlock_isem;
		}
	}
	xfs_iunlock(xip, XFS_ILOCK_EXCL);

	/*
	 * If we're writing the file then make sure to clear the
	 * setuid and setgid bits if the process is not being run
	 * by root.  This keeps people from modifying setuid and
	 * setgid binaries.
	 */

	if (((xip->i_d.di_mode & S_ISUID) ||
	    ((xip->i_d.di_mode & (S_ISGID | S_IXGRP)) ==
		(S_ISGID | S_IXGRP))) &&
	     !capable(CAP_FSETID)) {
		error = xfs_write_clear_setuid(xip);
		if (likely(!error))
			error = -remove_suid(file->f_dentry);
		if (unlikely(error)) {
			xfs_iunlock(xip, iolock);
			goto out_unlock_isem;
		}
	}

retry:
	/* We can write back this queue in page reclaim */
	current->backing_dev_info = mapping->backing_dev_info;

	if ((ioflags & IO_ISDIRECT)) {
		if (need_flush) {
			xfs_inval_cached_trace(io, pos, -1,
					ctooff(offtoct(pos)), -1);
			VOP_FLUSHINVAL_PAGES(vp, ctooff(offtoct(pos)),
					-1, FI_REMAPF_LOCKED);
		}

		if (need_isem) {
			/* demote the lock now the cached pages are gone */
			XFS_ILOCK_DEMOTE(mp, io, XFS_IOLOCK_EXCL);
			up(&inode->i_sem);

			iolock = XFS_IOLOCK_SHARED;
			locktype = VRWLOCK_WRITE_DIRECT;
			need_isem = 0;
		}

 		xfs_rw_enter_trace(XFS_DIOWR_ENTER, io, (void *)iovp, segs,
				*offset, ioflags);
		ret = generic_file_direct_write(iocb, iovp,
				&segs, pos, offset, count, ocount);

		/*
		 * direct-io write to a hole: fall through to buffered I/O
		 * for completing the rest of the request.
		 */
		if (ret >= 0 && ret != count) {
			XFS_STATS_ADD(xs_write_bytes, ret);

			pos += ret;
			count -= ret;

			need_isem = 1;
			ioflags &= ~IO_ISDIRECT;
			xfs_iunlock(xip, iolock);
			goto relock;
		}
	} else {
		xfs_rw_enter_trace(XFS_WRITE_ENTER, io, (void *)iovp, segs,
				*offset, ioflags);
		ret = generic_file_buffered_write(iocb, iovp, segs,
				pos, offset, count, ret);
	}

	current->backing_dev_info = NULL;

	if (ret == -EIOCBQUEUED)
		ret = wait_on_sync_kiocb(iocb);

	if ((ret == -ENOSPC) &&
	    DM_EVENT_ENABLED(vp->v_vfsp, xip, DM_EVENT_NOSPACE) &&
	    !(ioflags & IO_INVIS)) {

		xfs_rwunlock(bdp, locktype);
		error = XFS_SEND_NAMESP(xip->i_mount, DM_EVENT_NOSPACE, vp,
				DM_RIGHT_NULL, vp, DM_RIGHT_NULL, NULL, NULL,
				0, 0, 0); /* Delay flag intentionally  unused */
		if (error)
			goto out_unlock_isem;
		xfs_rwlock(bdp, locktype);
		pos = xip->i_d.di_size;
		goto retry;
	}

	if (*offset > xip->i_d.di_size) {
		xfs_ilock(xip, XFS_ILOCK_EXCL);
		if (*offset > xip->i_d.di_size) {
			xip->i_d.di_size = *offset;
			i_size_write(inode, *offset);
			xip->i_update_core = 1;
			xip->i_update_size = 1;
		}
		xfs_iunlock(xip, XFS_ILOCK_EXCL);
	}

	error = -ret;
	if (ret <= 0)
		goto out_unlock_internal;

	XFS_STATS_ADD(xs_write_bytes, ret);

	/* Handle various SYNC-type writes */
	if ((file->f_flags & O_SYNC) || IS_SYNC(inode)) {
		/*
		 * If we're treating this as O_DSYNC and we have not updated the
		 * size, force the log.
		 */
		if (!(mp->m_flags & XFS_MOUNT_OSYNCISOSYNC) &&
		    !(xip->i_update_size)) {
			xfs_inode_log_item_t	*iip = xip->i_itemp;

			/*
			 * If an allocation transaction occurred
			 * without extending the size, then we have to force
			 * the log up the proper point to ensure that the
			 * allocation is permanent.  We can't count on
			 * the fact that buffered writes lock out direct I/O
			 * writes - the direct I/O write could have extended
			 * the size nontransactionally, then finished before
			 * we started.  xfs_write_file will think that the file
			 * didn't grow but the update isn't safe unless the
			 * size change is logged.
			 *
			 * Force the log if we've committed a transaction
			 * against the inode or if someone else has and
			 * the commit record hasn't gone to disk (e.g.
			 * the inode is pinned).  This guarantees that
			 * all changes affecting the inode are permanent
			 * when we return.
			 */
			if (iip && iip->ili_last_lsn) {
				xfs_log_force(mp, iip->ili_last_lsn,
						XFS_LOG_FORCE | XFS_LOG_SYNC);
			} else if (xfs_ipincount(xip) > 0) {
				xfs_log_force(mp, (xfs_lsn_t)0,
						XFS_LOG_FORCE | XFS_LOG_SYNC);
			}

		} else {
			xfs_trans_t	*tp;

			/*
			 * O_SYNC or O_DSYNC _with_ a size update are handled
			 * the same way.
			 *
			 * If the write was synchronous then we need to make
			 * sure that the inode modification time is permanent.
			 * We'll have updated the timestamp above, so here
			 * we use a synchronous transaction to log the inode.
			 * It's not fast, but it's necessary.
			 *
			 * If this a dsync write and the size got changed
			 * non-transactionally, then we need to ensure that
			 * the size change gets logged in a synchronous
			 * transaction.
			 */

			tp = xfs_trans_alloc(mp, XFS_TRANS_WRITE_SYNC);
			if ((error = xfs_trans_reserve(tp, 0,
						      XFS_SWRITE_LOG_RES(mp),
						      0, 0, 0))) {
				/* Transaction reserve failed */
				xfs_trans_cancel(tp, 0);
			} else {
				/* Transaction reserve successful */
				xfs_ilock(xip, XFS_ILOCK_EXCL);
				xfs_trans_ijoin(tp, xip, XFS_ILOCK_EXCL);
				xfs_trans_ihold(tp, xip);
				xfs_trans_log_inode(tp, xip, XFS_ILOG_CORE);
				xfs_trans_set_sync(tp);
				error = xfs_trans_commit(tp, 0, NULL);
				xfs_iunlock(xip, XFS_ILOCK_EXCL);
				if (error)
					goto out_unlock_internal;
			}
		}
	
		xfs_rwunlock(bdp, locktype);
		if (need_isem)
			up(&inode->i_sem);

		error = sync_page_range(inode, mapping, pos, ret);
		if (!error)
			error = ret;
		return error;
	}

 out_unlock_internal:
	xfs_rwunlock(bdp, locktype);
 out_unlock_isem:
	if (need_isem)
		up(&inode->i_sem);
	return -error;
}
Exemple #4
0
/* FIXME: Ugliest function of all in LFS, need I say more? */
static ssize_t lfs_file_write( 	struct file *file, const char __user *buf, 	
				size_t count, loff_t *ppos)
{	
	loff_t pos;
	struct page *page;
	ssize_t res, written, bytes;
	struct inode *inode = file->f_dentry->d_inode;
	struct super_block *sb = inode->i_sb;
	struct segment *segp = LFS_SBI(sb)->s_curr;

	//dprintk("lfs_file_write called for %lu at pos %Lu\n", inode->i_ino, *ppos);
	if(file->f_flags & O_DIRECT) {
		dprintk("The file is requesting direct IO\n");
		return -EINVAL;
	}

	if (unlikely(count < 0 ))
		return -EINVAL;
	if (unlikely(!access_ok(VERIFY_READ, buf, count)))
		return -EFAULT;

	//down(&inode->i_sem);	/* lock the file */
	mutex_lock(&inode->i_mutex); //BrechREiZ: We need this for Kernel 2.6.17
	lfs_lock(sb);
		
	pos = *ppos;
	res = generic_write_checks(file, &pos, &count, 0);
	if (res)
		goto out;
	if(count == 0)
		goto out;
	
	res = remove_suid(file->f_dentry);
	if(res)
		goto out;
	//inode_update_time(inode, 1);	/* update mtime and ctime */
	file_update_time(inode); //BrechREiZ: We need this for Kernel 2.6.17

	written = 0;
	do {
		long offset;
		size_t copied;
		int i, siblock, eiblock, boffset;
		sector_t block;
				
		offset = (segp->offset % BUF_IN_PAGE) * LFS_BSIZE; 
		offset += pos & (LFS_BSIZE - 1); /* within block */
		bytes = PAGE_CACHE_SIZE - offset; /* number of bytes written
						     in this iteration */
		invalidate_old_page(inode, pos);

		if (bytes > count) 
			bytes = count;
		
		//dprintk("1:segp->start=%Lu,segp->offset=%d,segp->end=%Lu,offset=%lu,bytes=%d\n", segp->start, segp->offset, segp->end,offset,bytes);
		
		siblock = pos >> LFS_BSIZE_BITS;
		eiblock = (pos + bytes - 1) >> LFS_BSIZE_BITS;

		//dprintk("writing %d bytes at offset %ld (pos = %Lu)\n", bytes, offset, pos);
		//dprintk("siblock = %d, eiblock = %d\n", siblock, eiblock);
		

		/*
		 * Bring in the user page that we will copy from _first_.
		 * Otherwise there's a nasty deadlock on copying from the
		 * same page as we're writing to, without it being marked
		 * up-to-date.
		 */
		fault_in_pages_readable(buf, bytes);
		page = get_seg_page(segp);
		if (!page) {
			res = -ENOMEM;
			break;
		}

		/* fill the page with current inode blocks if any */
		boffset = offset / LFS_BSIZE;;
		for(i = siblock; i <= eiblock; ++i, ++boffset) {
			struct buffer_head *bh;
			//dprintk("Asking for block %d\n", i);
			bh = lfs_read_block(inode, i);
			if(!bh) /* new block */
				break;
			//dprintk("boffset = %d\n", boffset);
			memcpy(page_address(page) + LFS_BSIZE * boffset, bh->b_data, LFS_BSIZE);
			brelse(bh);
		}

		copied = __copy_from_user(page_address(page) + offset, buf, bytes);
		flush_dcache_page(page);

		block = segp->start + segp->offset;
		for(i = siblock;i <= eiblock; ++i, ++block)
			segsum_update_finfo(segp, inode->i_ino, i, block);

		block = segp->start + segp->offset;
		segp->offset += (bytes  - 1)/LFS_BSIZE + 1;
		//dprintk("2:segp->start=%Lu,segp->offset=%d,segp->end=%Lu,offset=%lu,bytes=%d\n",
		//segp->start, segp->offset, segp->end,offset,bytes);
		BUG_ON(segp->start + segp->offset > segp->end);
		if(segp->start + segp->offset == segp->end) {
			dprintk("allocating new segment\n");
			/* This also is going to write the previous segment */
			segment_allocate_new(inode->i_sb, segp, segp->start + segp->offset);
			segp = LFS_SBI(sb)->s_curr;
		}

		/* update the inode */
		for(i = siblock;i <= eiblock; ++i, ++block)
			update_inode(inode, i, block);
		//dprintk("start=%Lu,offset=%d,end=%Lu\n", segp->start, segp->offset, segp->end);
		segusetbl_add_livebytes(sb, segp->segnum, bytes);
		
		written += bytes;
		buf += bytes;
		pos += bytes;
		count -= bytes;
	} while(count);

	*ppos = pos;
	if(pos > inode->i_size)
		i_size_write(inode, pos);
	if(written)
		mark_inode_dirty(inode);
	
	lfs_unlock(sb);
	//up(&inode->i_sem);
	mutex_unlock(&inode->i_mutex); //BrechREiZ: and unlocking...
	return written ? written : res;
out:
	lfs_unlock(sb);
	//up(&inode->i_sem);
	mutex_unlock(&inode->i_mutex); //BrechREiZ: and unlocking...
	return res; 
}
Exemple #5
0
ssize_t				/* bytes written, or (-) error */
xfs_write(
	bhv_desc_t		*bdp,
	struct kiocb		*iocb,
	const struct iovec	*iovp,
	unsigned int		nsegs,
	loff_t			*offset,
	int			ioflags,
	cred_t			*credp)
{
	struct file		*file = iocb->ki_filp;
	struct address_space	*mapping = file->f_mapping;
	struct inode		*inode = mapping->host;
	unsigned long		segs = nsegs;
	xfs_inode_t		*xip;
	xfs_mount_t		*mp;
	ssize_t			ret = 0, error = 0;
	xfs_fsize_t		isize, new_size;
	xfs_iocore_t		*io;
	bhv_vnode_t		*vp;
	unsigned long		seg;
	int			iolock;
	int			eventsent = 0;
	bhv_vrwlock_t		locktype;
	size_t			ocount = 0, count;
	loff_t			pos;
	int			need_i_mutex = 1, need_flush = 0;

	XFS_STATS_INC(xs_write_calls);

	vp = BHV_TO_VNODE(bdp);
	xip = XFS_BHVTOI(bdp);

	for (seg = 0; seg < segs; seg++) {
		const struct iovec *iv = &iovp[seg];

		/*
		 * If any segment has a negative length, or the cumulative
		 * length ever wraps negative then return -EINVAL.
		 */
		ocount += iv->iov_len;
		if (unlikely((ssize_t)(ocount|iv->iov_len) < 0))
			return -EINVAL;
		if (access_ok(VERIFY_READ, iv->iov_base, iv->iov_len))
			continue;
		if (seg == 0)
			return -EFAULT;
		segs = seg;
		ocount -= iv->iov_len;  /* This segment is no good */
		break;
	}

	count = ocount;
	pos = *offset;

	if (count == 0)
		return 0;

	io = &xip->i_iocore;
	mp = io->io_mount;

	vfs_wait_for_freeze(vp->v_vfsp, SB_FREEZE_WRITE);

	if (XFS_FORCED_SHUTDOWN(mp))
		return -EIO;

	if (ioflags & IO_ISDIRECT) {
		xfs_buftarg_t	*target =
			(xip->i_d.di_flags & XFS_DIFLAG_REALTIME) ?
				mp->m_rtdev_targp : mp->m_ddev_targp;

		if ((pos & target->bt_smask) || (count & target->bt_smask))
			return XFS_ERROR(-EINVAL);

		if (!VN_CACHED(vp) && pos < i_size_read(inode))
			need_i_mutex = 0;

		if (VN_CACHED(vp))
			need_flush = 1;
	}

relock:
	if (need_i_mutex) {
		iolock = XFS_IOLOCK_EXCL;
		locktype = VRWLOCK_WRITE;

		mutex_lock(&inode->i_mutex);
	} else {
		iolock = XFS_IOLOCK_SHARED;
		locktype = VRWLOCK_WRITE_DIRECT;
	}

	xfs_ilock(xip, XFS_ILOCK_EXCL|iolock);

	isize = i_size_read(inode);

	if (file->f_flags & O_APPEND)
		*offset = isize;

start:
	error = -generic_write_checks(file, &pos, &count,
					S_ISBLK(inode->i_mode));
	if (error) {
		xfs_iunlock(xip, XFS_ILOCK_EXCL|iolock);
		goto out_unlock_mutex;
	}

	new_size = pos + count;
	if (new_size > isize)
		io->io_new_size = new_size;

	if ((DM_EVENT_ENABLED(vp->v_vfsp, xip, DM_EVENT_WRITE) &&
	    !(ioflags & IO_INVIS) && !eventsent)) {
		loff_t		savedsize = pos;
		int		dmflags = FILP_DELAY_FLAG(file);

		if (need_i_mutex)
			dmflags |= DM_FLAGS_IMUX;

		xfs_iunlock(xip, XFS_ILOCK_EXCL);
		error = XFS_SEND_DATA(xip->i_mount, DM_EVENT_WRITE, vp,
				      pos, count,
				      dmflags, &locktype);
		if (error) {
			xfs_iunlock(xip, iolock);
			goto out_unlock_mutex;
		}
		xfs_ilock(xip, XFS_ILOCK_EXCL);
		eventsent = 1;

		/*
		 * The iolock was dropped and reacquired in XFS_SEND_DATA
		 * so we have to recheck the size when appending.
		 * We will only "goto start;" once, since having sent the
		 * event prevents another call to XFS_SEND_DATA, which is
		 * what allows the size to change in the first place.
		 */
		if ((file->f_flags & O_APPEND) && savedsize != isize) {
			pos = isize = xip->i_d.di_size;
			goto start;
		}
	}

	if (likely(!(ioflags & IO_INVIS))) {
		file_update_time(file);
		xfs_ichgtime_fast(xip, inode,
				  XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
	}

	/*
	 * If the offset is beyond the size of the file, we have a couple
	 * of things to do. First, if there is already space allocated
	 * we need to either create holes or zero the disk or ...
	 *
	 * If there is a page where the previous size lands, we need
	 * to zero it out up to the new size.
	 */

	if (pos > isize) {
		error = xfs_zero_eof(BHV_TO_VNODE(bdp), io, pos, isize);
		if (error) {
			xfs_iunlock(xip, XFS_ILOCK_EXCL|iolock);
			goto out_unlock_mutex;
		}
	}
	xfs_iunlock(xip, XFS_ILOCK_EXCL);

	/*
	 * If we're writing the file then make sure to clear the
	 * setuid and setgid bits if the process is not being run
	 * by root.  This keeps people from modifying setuid and
	 * setgid binaries.
	 */

	if (((xip->i_d.di_mode & S_ISUID) ||
	    ((xip->i_d.di_mode & (S_ISGID | S_IXGRP)) ==
		(S_ISGID | S_IXGRP))) &&
	     !capable(CAP_FSETID)) {
		error = xfs_write_clear_setuid(xip);
		if (likely(!error))
			error = -remove_suid(file->f_path.dentry);
		if (unlikely(error)) {
			xfs_iunlock(xip, iolock);
			goto out_unlock_mutex;
		}
	}

retry:
	/* We can write back this queue in page reclaim */
	current->backing_dev_info = mapping->backing_dev_info;

	if ((ioflags & IO_ISDIRECT)) {
		if (need_flush) {
			xfs_inval_cached_trace(io, pos, -1,
					ctooff(offtoct(pos)), -1);
			bhv_vop_flushinval_pages(vp, ctooff(offtoct(pos)),
					-1, FI_REMAPF_LOCKED);
		}

		if (need_i_mutex) {
			/* demote the lock now the cached pages are gone */
			XFS_ILOCK_DEMOTE(mp, io, XFS_IOLOCK_EXCL);
			mutex_unlock(&inode->i_mutex);

			iolock = XFS_IOLOCK_SHARED;
			locktype = VRWLOCK_WRITE_DIRECT;
			need_i_mutex = 0;
		}

 		xfs_rw_enter_trace(XFS_DIOWR_ENTER, io, (void *)iovp, segs,
				*offset, ioflags);
		ret = generic_file_direct_write(iocb, iovp,
				&segs, pos, offset, count, ocount);

		/*
		 * direct-io write to a hole: fall through to buffered I/O
		 * for completing the rest of the request.
		 */
		if (ret >= 0 && ret != count) {
			XFS_STATS_ADD(xs_write_bytes, ret);

			pos += ret;
			count -= ret;

			need_i_mutex = 1;
			ioflags &= ~IO_ISDIRECT;
			xfs_iunlock(xip, iolock);
			goto relock;
		}
	} else {
		xfs_rw_enter_trace(XFS_WRITE_ENTER, io, (void *)iovp, segs,
				*offset, ioflags);
		ret = generic_file_buffered_write(iocb, iovp, segs,
				pos, offset, count, ret);
	}

	current->backing_dev_info = NULL;

	if (ret == -EIOCBQUEUED && !(ioflags & IO_ISAIO))
		ret = wait_on_sync_kiocb(iocb);

	if ((ret == -ENOSPC) &&
	    DM_EVENT_ENABLED(vp->v_vfsp, xip, DM_EVENT_NOSPACE) &&
	    !(ioflags & IO_INVIS)) {

		xfs_rwunlock(bdp, locktype);
		if (need_i_mutex)
			mutex_unlock(&inode->i_mutex);
		error = XFS_SEND_NAMESP(xip->i_mount, DM_EVENT_NOSPACE, vp,
				DM_RIGHT_NULL, vp, DM_RIGHT_NULL, NULL, NULL,
				0, 0, 0); /* Delay flag intentionally  unused */
		if (error)
			goto out_nounlocks;
		if (need_i_mutex)
			mutex_lock(&inode->i_mutex);
		xfs_rwlock(bdp, locktype);
		pos = xip->i_d.di_size;
		ret = 0;
		goto retry;
	}

	isize = i_size_read(inode);
	if (unlikely(ret < 0 && ret != -EFAULT && *offset > isize))
		*offset = isize;

	if (*offset > xip->i_d.di_size) {
		xfs_ilock(xip, XFS_ILOCK_EXCL);
		if (*offset > xip->i_d.di_size) {
			xip->i_d.di_size = *offset;
			i_size_write(inode, *offset);
			xip->i_update_core = 1;
			xip->i_update_size = 1;
		}
		xfs_iunlock(xip, XFS_ILOCK_EXCL);
	}

	error = -ret;
	if (ret <= 0)
		goto out_unlock_internal;

	XFS_STATS_ADD(xs_write_bytes, ret);

	/* Handle various SYNC-type writes */
	if ((file->f_flags & O_SYNC) || IS_SYNC(inode)) {
		error = xfs_write_sync_logforce(mp, xip);
		if (error)
			goto out_unlock_internal;

		xfs_rwunlock(bdp, locktype);
		if (need_i_mutex)
			mutex_unlock(&inode->i_mutex);

		error = sync_page_range(inode, mapping, pos, ret);
		if (!error)
			error = ret;
		return error;
	}

 out_unlock_internal:
	xfs_rwunlock(bdp, locktype);
 out_unlock_mutex:
	if (need_i_mutex)
		mutex_unlock(&inode->i_mutex);
 out_nounlocks:
	return -error;
}
Exemple #6
0
ssize_t				/* bytes written, or (-) error */
xfs_write(
	bhv_desc_t      *bdp,
	uio_t		*uio,
	int		ioflag,
	cred_t          *credp)
{
	xfs_inode_t	*xip;
	xfs_mount_t	*mp;
	ssize_t		ret = 0;
	int		error = 0;
	xfs_fsize_t     isize, new_size;
	xfs_fsize_t	n, limit;
	xfs_fsize_t	size;
	xfs_iocore_t    *io;
	xfs_vnode_t	*vp;
	int		iolock;
	//int		eventsent = 0;
	vrwlock_t	locktype;
	xfs_off_t	offset_c;
	xfs_off_t	*offset;
	xfs_off_t	pos;

	XFS_STATS_INC(xs_write_calls);

	vp = BHV_TO_VNODE(bdp);
	xip = XFS_BHVTOI(bdp);

	io = &xip->i_iocore;
	mp = io->io_mount;

	if (XFS_FORCED_SHUTDOWN(xip->i_mount)) {
		return EIO;
	}

	size = uio->uio_resid;
	pos = offset_c = uio->uio_offset;
	offset = &offset_c;

	if (unlikely(ioflag & IO_ISDIRECT)) {
		if (((__psint_t)buf & BBMASK) ||
		    (*offset & mp->m_blockmask) ||
		    (size  & mp->m_blockmask)) {
			return EINVAL;
		}
		iolock = XFS_IOLOCK_SHARED;
		locktype = VRWLOCK_WRITE_DIRECT;
	} else {
		if (io->io_flags & XFS_IOCORE_RT)
			return EINVAL;
		iolock = XFS_IOLOCK_EXCL;
		locktype = VRWLOCK_WRITE;
	}

	iolock = XFS_IOLOCK_EXCL;
	locktype = VRWLOCK_WRITE;

	xfs_ilock(xip, XFS_ILOCK_EXCL|iolock);

	isize = xip->i_d.di_size;
	limit = XFS_MAXIOFFSET(mp);

	if (ioflag & O_APPEND)
		*offset = isize;

//start:
	n = limit - *offset;
	if (n <= 0) {
		xfs_iunlock(xip, XFS_ILOCK_EXCL|iolock);
		return EFBIG;
	}
	if (n < size)
		size = n;

	new_size = *offset + size;
	if (new_size > isize) {
		io->io_new_size = new_size;
	}

#ifdef RMC
	/* probably be a long time before if ever that we do dmapi */
	if ((DM_EVENT_ENABLED(vp->v_vfsp, xip, DM_EVENT_WRITE) &&
	    !(ioflags & IO_INVIS) && !eventsent)) {
		loff_t		savedsize = *offset;
		int dmflags = FILP_DELAY_FLAG(file) | DM_SEM_FLAG_RD(ioflags);

		xfs_iunlock(xip, XFS_ILOCK_EXCL);
		error = XFS_SEND_DATA(xip->i_mount, DM_EVENT_WRITE, vp,
				      *offset, size,
				      dmflags, &locktype);
		if (error) {
			if (iolock) xfs_iunlock(xip, iolock);
			return -error;
		}
		xfs_ilock(xip, XFS_ILOCK_EXCL);
		eventsent = 1;

		/*
		 * The iolock was dropped and reaquired in XFS_SEND_DATA
		 * so we have to recheck the size when appending.
		 * We will only "goto start;" once, since having sent the
		 * event prevents another call to XFS_SEND_DATA, which is
		 * what allows the size to change in the first place.
		 */
		if ((file->f_flags & O_APPEND) &&
		    savedsize != xip->i_d.di_size) {
			*offset = isize = xip->i_d.di_size;
			goto start;
		}
	}
#endif

	/*
	 * If the offset is beyond the size of the file, we have a couple
	 * of things to do. First, if there is already space allocated
	 * we need to either create holes or zero the disk or ...
	 *
	 * If there is a page where the previous size lands, we need
	 * to zero it out up to the new size.
	 */

	if (!(ioflag & IO_ISDIRECT) && (*offset > isize && isize)) {
		error = xfs_zero_eof(BHV_TO_VNODE(bdp), io, *offset,
			isize, *offset + size);
		if (error) {
			xfs_iunlock(xip, XFS_ILOCK_EXCL|iolock);
			return(-error);
		}
	}
	xfs_iunlock(xip, XFS_ILOCK_EXCL);

#if 0
	/*
	 * If we're writing the file then make sure to clear the
	 * setuid and setgid bits if the process is not being run
	 * by root.  This keeps people from modifying setuid and
	 * setgid binaries.
	 */

	if (((xip->i_d.di_mode & S_ISUID) ||
	    ((xip->i_d.di_mode & (S_ISGID | S_IXGRP)) ==
		(S_ISGID | S_IXGRP))) &&
	     !capable(CAP_FSETID)) {
		error = xfs_write_clear_setuid(xip);
		if (likely(!error))
			error = -remove_suid(file->f_dentry);
		if (unlikely(error)) {
			xfs_iunlock(xip, iolock);
			goto out_unlock_mutex;
		}
	}
#endif


//retry:
	if (unlikely(ioflag & IO_ISDIRECT)) {

#ifdef RMC
		xfs_off_t	pos = *offset;
		struct address_space *mapping = file->f_dentry->d_inode->i_mapping;
		struct inode    *inode = mapping->host;

		ret = precheck_file_write(file, inode, &size,  &pos);
		if (ret || size == 0)
			goto error;

		xfs_inval_cached_pages(vp, io, pos, 1, 1);
		inode->i_ctime = inode->i_mtime = CURRENT_TIME;
		/* mark_inode_dirty_sync(inode); - we do this later */

		xfs_rw_enter_trace(XFS_DIOWR_ENTER, io, buf, size, pos, ioflags);
		ret = generic_file_direct_IO(WRITE, file, (char *)buf, size, pos);
		xfs_inval_cached_pages(vp, io, pos, 1, 1);
		if (ret > 0)
			*offset += ret;
#endif
	} else {
		xfs_rw_enter_trace(XFS_WRITE_ENTER, io, buf, size, *offset, ioflags);
		ret = xfs_write_file(xip,uio,ioflag);
	}

	xfs_ichgtime(xip, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);


//error:
	if (ret <= 0) {
		if (iolock)
			xfs_rwunlock(bdp, locktype);
		return ret;
	}

	XFS_STATS_ADD(xs_write_bytes, ret);

	if (*offset > xip->i_d.di_size) {
		xfs_ilock(xip, XFS_ILOCK_EXCL);
		if (*offset > xip->i_d.di_size) {
			printf("xfs_write look at doing more here %s:%d\n",__FILE__,__LINE__);
#ifdef RMC
			struct inode	*inode = LINVFS_GET_IP(vp);
			i_size_write(inode, *offset);
			mark_inode_dirty_sync(inode);
#endif

			xip->i_d.di_size = *offset;
			xip->i_update_core = 1;
			xip->i_update_size = 1;
		}
		xfs_iunlock(xip, XFS_ILOCK_EXCL);
	}

	/* Handle various SYNC-type writes */
#if 0
//	if ((file->f_flags & O_SYNC) || IS_SYNC(inode)) {
#endif
	if (ioflag & IO_SYNC) {
		/*
		 * If we're treating this as O_DSYNC and we have not updated the
		 * size, force the log.
		 */
		if (!(mp->m_flags & XFS_MOUNT_OSYNCISOSYNC) &&
		    !(xip->i_update_size)) {
			xfs_inode_log_item_t	*iip = xip->i_itemp;

			/*
			 * If an allocation transaction occurred
			 * without extending the size, then we have to force
			 * the log up the proper point to ensure that the
			 * allocation is permanent.  We can't count on
			 * the fact that buffered writes lock out direct I/O
			 * writes - the direct I/O write could have extended
			 * the size nontransactionally, then finished before
			 * we started.  xfs_write_file will think that the file
			 * didn't grow but the update isn't safe unless the
			 * size change is logged.
			 *
			 * Force the log if we've committed a transaction
			 * against the inode or if someone else has and
			 * the commit record hasn't gone to disk (e.g.
			 * the inode is pinned).  This guarantees that
			 * all changes affecting the inode are permanent
			 * when we return.
			 */
			if (iip && iip->ili_last_lsn) {
				xfs_log_force(mp, iip->ili_last_lsn,
						XFS_LOG_FORCE | XFS_LOG_SYNC);
			} else if (xfs_ipincount(xip) > 0) {
				xfs_log_force(mp, (xfs_lsn_t)0,
						XFS_LOG_FORCE | XFS_LOG_SYNC);
			}

		} else {
			xfs_trans_t	*tp;

			/*
			 * O_SYNC or O_DSYNC _with_ a size update are handled
			 * the same way.
			 *
			 * If the write was synchronous then we need to make
			 * sure that the inode modification time is permanent.
			 * We'll have updated the timestamp above, so here
			 * we use a synchronous transaction to log the inode.
			 * It's not fast, but it's necessary.
			 *
			 * If this a dsync write and the size got changed
			 * non-transactionally, then we need to ensure that
			 * the size change gets logged in a synchronous
			 * transaction.
			 */

			tp = xfs_trans_alloc(mp, XFS_TRANS_WRITE_SYNC);
			if ((error = xfs_trans_reserve(tp, 0,
						      XFS_SWRITE_LOG_RES(mp),
						      0, 0, 0))) {
				/* Transaction reserve failed */
				xfs_trans_cancel(tp, 0);
			} else {
				/* Transaction reserve successful */
				xfs_ilock(xip, XFS_ILOCK_EXCL);
				xfs_trans_ijoin(tp, xip, XFS_ILOCK_EXCL);
				xfs_trans_ihold(tp, xip);
				xfs_trans_log_inode(tp, xip, XFS_ILOG_CORE);
				xfs_trans_set_sync(tp);
				error = xfs_trans_commit(tp, 0, NULL);
				xfs_iunlock(xip, XFS_ILOCK_EXCL);
			}
			if (error)
				goto out_unlock_internal;
		}

		xfs_rwunlock(bdp, locktype);
		return ret;

	} /* (ioflags & O_SYNC) */

out_unlock_internal:
	xfs_rwunlock(bdp, locktype);
#if 0
out_unlock_mutex:
	if (need_i_mutex)
		mutex_unlock(&inode->i_mutex);
#endif
 //out_nounlocks:
	return -error;
}
Exemple #7
0
ssize_t				/* bytes written, or (-) error */
xfs_write(
	struct xfs_inode	*xip,
	struct kiocb		*iocb,
	const struct iovec	*iovp,
	unsigned int		nsegs,
	loff_t			*offset,
	int			ioflags)
{
	struct file		*file = iocb->ki_filp;
	struct address_space	*mapping = file->f_mapping;
	struct inode		*inode = mapping->host;
	bhv_vnode_t		*vp = XFS_ITOV(xip);
	unsigned long		segs = nsegs;
	xfs_mount_t		*mp;
	ssize_t			ret = 0, error = 0;
	xfs_fsize_t		isize, new_size;
	int			iolock;
	int			eventsent = 0;
	bhv_vrwlock_t		locktype;
	size_t			ocount = 0, count;
	loff_t			pos;
	int			need_i_mutex;

	XFS_STATS_INC(xs_write_calls);

	error = generic_segment_checks(iovp, &segs, &ocount, VERIFY_READ);
	if (error)
		return error;

	count = ocount;
	pos = *offset;

	if (count == 0)
		return 0;

	mp = xip->i_mount;

	xfs_wait_for_freeze(mp, SB_FREEZE_WRITE);

	if (XFS_FORCED_SHUTDOWN(mp))
		return -EIO;

relock:
	if (ioflags & IO_ISDIRECT) {
		iolock = XFS_IOLOCK_SHARED;
		locktype = VRWLOCK_WRITE_DIRECT;
		need_i_mutex = 0;
	} else {
		iolock = XFS_IOLOCK_EXCL;
		locktype = VRWLOCK_WRITE;
		need_i_mutex = 1;
		mutex_lock(&inode->i_mutex);
	}

	xfs_ilock(xip, XFS_ILOCK_EXCL|iolock);

start:
	error = -generic_write_checks(file, &pos, &count,
					S_ISBLK(inode->i_mode));
	if (error) {
		xfs_iunlock(xip, XFS_ILOCK_EXCL|iolock);
		goto out_unlock_mutex;
	}

	if ((DM_EVENT_ENABLED(xip, DM_EVENT_WRITE) &&
	    !(ioflags & IO_INVIS) && !eventsent)) {
		int		dmflags = FILP_DELAY_FLAG(file);

		if (need_i_mutex)
			dmflags |= DM_FLAGS_IMUX;

		xfs_iunlock(xip, XFS_ILOCK_EXCL);
		error = XFS_SEND_DATA(xip->i_mount, DM_EVENT_WRITE, vp,
				      pos, count,
				      dmflags, &locktype);
		if (error) {
			goto out_unlock_internal;
		}
		xfs_ilock(xip, XFS_ILOCK_EXCL);
		eventsent = 1;

		/*
		 * The iolock was dropped and reacquired in XFS_SEND_DATA
		 * so we have to recheck the size when appending.
		 * We will only "goto start;" once, since having sent the
		 * event prevents another call to XFS_SEND_DATA, which is
		 * what allows the size to change in the first place.
		 */
		if ((file->f_flags & O_APPEND) && pos != xip->i_size)
			goto start;
	}

	if (ioflags & IO_ISDIRECT) {
		xfs_buftarg_t	*target =
			XFS_IS_REALTIME_INODE(xip) ?
				mp->m_rtdev_targp : mp->m_ddev_targp;

		if ((pos & target->bt_smask) || (count & target->bt_smask)) {
			xfs_iunlock(xip, XFS_ILOCK_EXCL|iolock);
			return XFS_ERROR(-EINVAL);
		}

		if (!need_i_mutex && (VN_CACHED(vp) || pos > xip->i_size)) {
			xfs_iunlock(xip, XFS_ILOCK_EXCL|iolock);
			iolock = XFS_IOLOCK_EXCL;
			locktype = VRWLOCK_WRITE;
			need_i_mutex = 1;
			mutex_lock(&inode->i_mutex);
			xfs_ilock(xip, XFS_ILOCK_EXCL|iolock);
			goto start;
		}
	}

	new_size = pos + count;
	if (new_size > xip->i_size)
		xip->i_new_size = new_size;

	if (likely(!(ioflags & IO_INVIS))) {
		file_update_time(file);
		xfs_ichgtime_fast(xip, inode,
				  XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
	}

	/*
	 * If the offset is beyond the size of the file, we have a couple
	 * of things to do. First, if there is already space allocated
	 * we need to either create holes or zero the disk or ...
	 *
	 * If there is a page where the previous size lands, we need
	 * to zero it out up to the new size.
	 */

	if (pos > xip->i_size) {
		error = xfs_zero_eof(xip, pos, xip->i_size);
		if (error) {
			xfs_iunlock(xip, XFS_ILOCK_EXCL);
			goto out_unlock_internal;
		}
	}
	xfs_iunlock(xip, XFS_ILOCK_EXCL);

	/*
	 * If we're writing the file then make sure to clear the
	 * setuid and setgid bits if the process is not being run
	 * by root.  This keeps people from modifying setuid and
	 * setgid binaries.
	 */

	if (((xip->i_d.di_mode & S_ISUID) ||
	    ((xip->i_d.di_mode & (S_ISGID | S_IXGRP)) ==
		(S_ISGID | S_IXGRP))) &&
	     !capable(CAP_FSETID)) {
		error = xfs_write_clear_setuid(xip);
		if (likely(!error))
			error = -remove_suid(file->f_path.dentry);
		if (unlikely(error)) {
			goto out_unlock_internal;
		}
	}

retry:
	/* We can write back this queue in page reclaim */
	current->backing_dev_info = mapping->backing_dev_info;

	if ((ioflags & IO_ISDIRECT)) {
		if (VN_CACHED(vp)) {
			WARN_ON(need_i_mutex == 0);
			xfs_inval_cached_trace(xip, pos, -1,
					(pos & PAGE_CACHE_MASK), -1);
			error = xfs_flushinval_pages(xip,
					(pos & PAGE_CACHE_MASK),
					-1, FI_REMAPF_LOCKED);
			if (error)
				goto out_unlock_internal;
		}

		if (need_i_mutex) {
			/* demote the lock now the cached pages are gone */
			xfs_ilock_demote(xip, XFS_IOLOCK_EXCL);
			mutex_unlock(&inode->i_mutex);

			iolock = XFS_IOLOCK_SHARED;
			locktype = VRWLOCK_WRITE_DIRECT;
			need_i_mutex = 0;
		}

 		xfs_rw_enter_trace(XFS_DIOWR_ENTER, xip, (void *)iovp, segs,
				*offset, ioflags);
		ret = generic_file_direct_write(iocb, iovp,
				&segs, pos, offset, count, ocount);

		/*
		 * direct-io write to a hole: fall through to buffered I/O
		 * for completing the rest of the request.
		 */
		if (ret >= 0 && ret != count) {
			XFS_STATS_ADD(xs_write_bytes, ret);

			pos += ret;
			count -= ret;

			ioflags &= ~IO_ISDIRECT;
			xfs_iunlock(xip, iolock);
			goto relock;
		}
	} else {
		xfs_rw_enter_trace(XFS_WRITE_ENTER, xip, (void *)iovp, segs,
				*offset, ioflags);
		ret = generic_file_buffered_write(iocb, iovp, segs,
				pos, offset, count, ret);
	}

	current->backing_dev_info = NULL;

	if (ret == -EIOCBQUEUED && !(ioflags & IO_ISAIO))
		ret = wait_on_sync_kiocb(iocb);

	if (ret == -ENOSPC &&
	    DM_EVENT_ENABLED(xip, DM_EVENT_NOSPACE) && !(ioflags & IO_INVIS)) {
		xfs_rwunlock(xip, locktype);
		if (need_i_mutex)
			mutex_unlock(&inode->i_mutex);
		error = XFS_SEND_NAMESP(xip->i_mount, DM_EVENT_NOSPACE, vp,
				DM_RIGHT_NULL, vp, DM_RIGHT_NULL, NULL, NULL,
				0, 0, 0); /* Delay flag intentionally  unused */
		if (need_i_mutex)
			mutex_lock(&inode->i_mutex);
		xfs_rwlock(xip, locktype);
		if (error)
			goto out_unlock_internal;
		pos = xip->i_size;
		ret = 0;
		goto retry;
	}

	isize = i_size_read(inode);
	if (unlikely(ret < 0 && ret != -EFAULT && *offset > isize))
		*offset = isize;

	if (*offset > xip->i_size) {
		xfs_ilock(xip, XFS_ILOCK_EXCL);
		if (*offset > xip->i_size)
			xip->i_size = *offset;
		xfs_iunlock(xip, XFS_ILOCK_EXCL);
	}

	error = -ret;
	if (ret <= 0)
		goto out_unlock_internal;

	XFS_STATS_ADD(xs_write_bytes, ret);

	/* Handle various SYNC-type writes */
	if ((file->f_flags & O_SYNC) || IS_SYNC(inode)) {
		int error2;
		xfs_rwunlock(xip, locktype);
		if (need_i_mutex)
			mutex_unlock(&inode->i_mutex);
		error2 = sync_page_range(inode, mapping, pos, ret);
		if (!error)
			error = error2;
		if (need_i_mutex)
			mutex_lock(&inode->i_mutex);
		xfs_rwlock(xip, locktype);
		error2 = xfs_write_sync_logforce(mp, xip);
		if (!error)
			error = error2;
	}

 out_unlock_internal:
	if (xip->i_new_size) {
		xfs_ilock(xip, XFS_ILOCK_EXCL);
		xip->i_new_size = 0;
		/*
		 * If this was a direct or synchronous I/O that failed (such
		 * as ENOSPC) then part of the I/O may have been written to
		 * disk before the error occured.  In this case the on-disk
		 * file size may have been adjusted beyond the in-memory file
		 * size and now needs to be truncated back.
		 */
		if (xip->i_d.di_size > xip->i_size)
			xip->i_d.di_size = xip->i_size;
		xfs_iunlock(xip, XFS_ILOCK_EXCL);
	}
	xfs_rwunlock(xip, locktype);
 out_unlock_mutex:
	if (need_i_mutex)
		mutex_unlock(&inode->i_mutex);
	return -error;
}
Exemple #8
0
long ext4_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
{
	struct inode *inode = filp->f_dentry->d_inode;
	struct ext4_inode_info *ei = EXT4_I(inode);
	unsigned int flags;

	ext4_debug("cmd = %u, arg = %lu\n", cmd, arg);

	switch (cmd) {
	case EXT4_IOC_GETFLAGS:
		ext4_get_inode_flags(ei);
		flags = ei->i_flags & EXT4_FL_USER_VISIBLE;
		return put_user(flags, (int __user *) arg);
	case EXT4_IOC_SETFLAGS: {
		handle_t *handle = NULL;
		int err, migrate = 0;
		struct ext4_iloc iloc;
		unsigned int oldflags;
		unsigned int jflag;

		if (IS_RDONLY(inode))
			return -EROFS;

		if ((current->fsuid != inode->i_uid) && !capable(CAP_FOWNER))
			return -EACCES;

		if (get_user(flags, (int __user *) arg))
			return -EFAULT;

		flags = ext4_mask_flags(inode->i_mode, flags);

		mutex_lock(&inode->i_mutex);
		/* Is it quota file? Do not allow user to mess with it */
		if (IS_NOQUOTA(inode)) {
			mutex_unlock(&inode->i_mutex);
			return -EPERM;
		}
		oldflags = ei->i_flags;

		/* The JOURNAL_DATA flag is modifiable only by root */
		jflag = flags & EXT4_JOURNAL_DATA_FL;

		/*
		 * The IMMUTABLE and APPEND_ONLY flags can only be changed by
		 * the relevant capability.
		 *
		 * This test looks nicer. Thanks to Pauline Middelink
		 */
		if ((flags ^ oldflags) & (EXT4_APPEND_FL | EXT4_IMMUTABLE_FL)) {
			if (!capable(CAP_LINUX_IMMUTABLE)) {
				mutex_unlock(&inode->i_mutex);
				return -EPERM;
			}
		}

		/*
		 * The JOURNAL_DATA flag can only be changed by
		 * the relevant capability.
		 */
		if ((jflag ^ oldflags) & (EXT4_JOURNAL_DATA_FL)) {
 			if (!capable(CAP_SYS_RESOURCE)) {
 				mutex_unlock(&inode->i_mutex);
 				return -EPERM;
 			}
		}
		if (oldflags & EXT4_EXTENTS_FL) {
			/* We don't support clearning extent flags */
			if (!(flags & EXT4_EXTENTS_FL)) {
				err = -EOPNOTSUPP;
				goto flags_out;
			}
		} else if (flags & EXT4_EXTENTS_FL) {
			/* migrate the file */
			migrate = 1;
			flags &= ~EXT4_EXTENTS_FL;
		}

		if (flags & EXT4_EOFBLOCKS_FL) {
			/* we don't support adding EOFBLOCKS flag */
			if (!(oldflags & EXT4_EOFBLOCKS_FL)) {
				err = -EOPNOTSUPP;
				goto flags_out;
			}
		} else if (oldflags & EXT4_EOFBLOCKS_FL)
			ext4_truncate(inode);

		handle = ext4_journal_start(inode, 1);
		if (IS_ERR(handle)) {
			err = PTR_ERR(handle);
			goto flags_out;
		}
		if (IS_SYNC(inode))
			ext4_handle_sync(handle);
		err = ext4_reserve_inode_write(handle, inode, &iloc);
		if (err)
			goto flags_err;

		flags = flags & EXT4_FL_USER_MODIFIABLE;
		flags |= oldflags & ~EXT4_FL_USER_MODIFIABLE;
		ei->i_flags = flags;

		ext4_set_inode_flags(inode);
		inode->i_ctime = ext4_current_time(inode);

		err = ext4_mark_iloc_dirty(handle, inode, &iloc);
flags_err:
		ext4_journal_stop(handle);
		if (err)
			goto flags_out;

		if ((jflag ^ oldflags) & (EXT4_JOURNAL_DATA_FL))
			err = ext4_change_inode_journal_flag(inode, jflag);
		if (err)
			goto flags_out;
		if (migrate)
			err = ext4_ext_migrate(inode);
flags_out:
		mutex_unlock(&inode->i_mutex);
		return err;
	}
	case EXT4_IOC_GETVERSION:
	case EXT4_IOC_GETVERSION_OLD:
		return put_user(inode->i_generation, (int __user *) arg);
	case EXT4_IOC_SETVERSION:
	case EXT4_IOC_SETVERSION_OLD: {
		handle_t *handle;
		struct ext4_iloc iloc;
		__u32 generation;
		int err;

		if ((current->fsuid != inode->i_uid) && !capable(CAP_FOWNER))
			return -EPERM;
		if (IS_RDONLY(inode))
			return -EROFS;
		if (get_user(generation, (int __user *) arg))
			return -EFAULT;

		handle = ext4_journal_start(inode, 1);
		if (IS_ERR(handle))
			return PTR_ERR(handle);
		err = ext4_reserve_inode_write(handle, inode, &iloc);
		if (err == 0) {
			inode->i_ctime = ext4_current_time(inode);
			inode->i_generation = generation;
			err = ext4_mark_iloc_dirty(handle, inode, &iloc);
		}
		ext4_journal_stop(handle);
		return err;
	}
#ifdef CONFIG_JBD2_DEBUG
	case EXT4_IOC_WAIT_FOR_READONLY:
		/*
		 * This is racy - by the time we're woken up and running,
		 * the superblock could be released.  And the module could
		 * have been unloaded.  So sue me.
		 *
		 * Returns 1 if it slept, else zero.
		 */
		{
			struct super_block *sb = inode->i_sb;
			DECLARE_WAITQUEUE(wait, current);
			int ret = 0;

			set_current_state(TASK_INTERRUPTIBLE);
			add_wait_queue(&EXT4_SB(sb)->ro_wait_queue, &wait);
			if (timer_pending(&EXT4_SB(sb)->turn_ro_timer)) {
				schedule();
				ret = 1;
			}
			remove_wait_queue(&EXT4_SB(sb)->ro_wait_queue, &wait);
			return ret;
		}
#endif
	case EXT4_IOC_GROUP_EXTEND: {
		ext4_fsblk_t n_blocks_count;
		struct super_block *sb = inode->i_sb;
		int err, err2=0;

		if (!capable(CAP_SYS_RESOURCE))
			return -EPERM;

		if (get_user(n_blocks_count, (__u32 __user *)arg))
			return -EFAULT;

		if (IS_RDONLY(inode))
			return -EROFS;

		err = ext4_group_extend(sb, EXT4_SB(sb)->s_es, n_blocks_count);
		if (EXT4_SB(sb)->s_journal) {
			jbd2_journal_lock_updates(EXT4_SB(sb)->s_journal);
			err2 = jbd2_journal_flush(EXT4_SB(sb)->s_journal);
			jbd2_journal_unlock_updates(EXT4_SB(sb)->s_journal);
		}
		if (err == 0)
			err = err2;

		return err;
	}

	case EXT4_IOC_MOVE_EXT: {
		struct move_extent me;
		struct file *donor_filp;
		int err;

		if (!(filp->f_mode & FMODE_READ) ||
		    !(filp->f_mode & FMODE_WRITE))
			return -EBADF;

		if (copy_from_user(&me,
			(struct move_extent __user *)arg, sizeof(me)))
			return -EFAULT;
		me.moved_len = 0;

		donor_filp = fget(me.donor_fd);
		if (!donor_filp)
			return -EBADF;

		if (!(donor_filp->f_mode & FMODE_WRITE)) {
			err = -EBADF;
			goto mext_out;
		}

		if (IS_RDONLY(inode))
			return -EROFS;

		err = ext4_move_extents(filp, donor_filp, me.orig_start,
					me.donor_start, me.len, &me.moved_len);
		if (me.moved_len > 0)
			remove_suid(donor_filp->f_path.dentry);

		if (copy_to_user((struct move_extent __user *)arg,
				 &me, sizeof(me)))
			err = -EFAULT;
mext_out:
		fput(donor_filp);
		return err;
	}

	case EXT4_IOC_GROUP_ADD: {
		struct ext4_new_group_data input;
		struct super_block *sb = inode->i_sb;
		int err, err2=0;

		if (!capable(CAP_SYS_RESOURCE))
			return -EPERM;

		if (copy_from_user(&input, (struct ext4_new_group_input __user *)arg,
				sizeof(input)))
			return -EFAULT;

		if (IS_RDONLY(inode))
			return -EROFS;

		err = ext4_group_add(sb, &input);
		if (EXT4_SB(sb)->s_journal) {
			jbd2_journal_lock_updates(EXT4_SB(sb)->s_journal);
			err2 = jbd2_journal_flush(EXT4_SB(sb)->s_journal);
			jbd2_journal_unlock_updates(EXT4_SB(sb)->s_journal);
		}
		if (err == 0)
			err = err2;

		return err;
	}

	case EXT4_IOC_MIGRATE:
	{
		int err;
		if ((current->fsuid != inode->i_uid) && !capable(CAP_FOWNER))
			return -EACCES;

		if (IS_RDONLY(inode))
			return -EROFS;
		/*
		 * inode_mutex prevent write and truncate on the file.
		 * Read still goes through. We take i_data_sem in
		 * ext4_ext_swap_inode_data before we switch the
		 * inode format to prevent read.
		 */
		mutex_lock(&(inode->i_mutex));
		err = ext4_ext_migrate(inode);
		mutex_unlock(&(inode->i_mutex));
		return err;
	}

	case EXT4_IOC_ALLOC_DA_BLKS:
	{
		int err;
		if ((current->fsuid != inode->i_uid) && !capable(CAP_FOWNER))
			return -EACCES;

		if (IS_RDONLY(inode))
			return -EROFS;

		err = ext4_alloc_da_blocks(inode);
		return err;
	}

	default:
		return -ENOTTY;
	}
}