Donate to e Foundation | Murena handsets with /e/OS | Own a part of Murena! Learn more

Commit 65832940 authored by Linus Torvalds's avatar Linus Torvalds
Browse files

Merge branch 'for-linus' of git://oss.sgi.com/xfs/xfs 

* 'for-linus' of git://oss.sgi.com/xfs/xfs:
  xfs: don't warn on EAGAIN in inode reclaim
  xfs: ensure that sync updates the log tail correctly
parents 1c1ec9c0 f1d486a3

fs/xfs/linux-2.6/xfs_sync.c

+2 −2
Original line number Diff line number Diff line
@@ -820,10 +820,10 @@ xfs_reclaim_inode( 
	 * call into reclaim to find it in a clean state instead of waiting for

	 * it now. We also don't return errors here - if the error is transient

	 * then the next reclaim pass will flush the inode, and if the error

	 * is permanent then the next sync reclaim will relcaim the inode and

	 * is permanent then the next sync reclaim will reclaim the inode and

	 * pass on the error.

	 */

	if (error && !XFS_FORCED_SHUTDOWN(ip->i_mount)) {

	if (error && error != EAGAIN && !XFS_FORCED_SHUTDOWN(ip->i_mount)) {

		xfs_fs_cmn_err(CE_WARN, ip->i_mount,

			"inode 0x%llx background reclaim flush failed with %d",

			(long long)ip->i_ino, error);

fs/xfs/xfs_log.c

+26 −12
Original line number Diff line number Diff line
@@ -745,9 +745,16 @@ xfs_log_move_tail(xfs_mount_t *mp, 


/*

 * Determine if we have a transaction that has gone to disk

 * that needs to be covered. Log activity needs to be idle (no AIL and

 * nothing in the iclogs). And, we need to be in the right state indicating

 * something has gone out.

 * that needs to be covered. To begin the transition to the idle state

 * firstly the log needs to be idle (no AIL and nothing in the iclogs).

 * If we are then in a state where covering is needed, the caller is informed

 * that dummy transactions are required to move the log into the idle state.

 *

 * Because this is called as part of the sync process, we should also indicate

 * that dummy transactions should be issued in anything but the covered or

 * idle states. This ensures that the log tail is accurately reflected in

 * the log at the end of the sync, hence if a crash occurrs avoids replay

 * of transactions where the metadata is already on disk.

 */

int

xfs_log_need_covered(xfs_mount_t *mp)
@@ -759,17 +766,24 @@ xfs_log_need_covered(xfs_mount_t *mp) 
		return 0;



	spin_lock(&log->l_icloglock);

	if (((log->l_covered_state == XLOG_STATE_COVER_NEED) ||

		(log->l_covered_state == XLOG_STATE_COVER_NEED2))

			&& !xfs_trans_ail_tail(log->l_ailp)

			&& xlog_iclogs_empty(log)) {

	switch (log->l_covered_state) {

	case XLOG_STATE_COVER_DONE:

	case XLOG_STATE_COVER_DONE2:

	case XLOG_STATE_COVER_IDLE:

		break;

	case XLOG_STATE_COVER_NEED:

	case XLOG_STATE_COVER_NEED2:

		if (!xfs_trans_ail_tail(log->l_ailp) &&

		    xlog_iclogs_empty(log)) {

			if (log->l_covered_state == XLOG_STATE_COVER_NEED)

				log->l_covered_state = XLOG_STATE_COVER_DONE;

		else {

			ASSERT(log->l_covered_state == XLOG_STATE_COVER_NEED2);

			else

				log->l_covered_state = XLOG_STATE_COVER_DONE2;

		}

		/* FALLTHRU */

	default:

		needed = 1;

		break;

	}

	spin_unlock(&log->l_icloglock);

	return needed;