f2fs: guarantee journalled quota data by checkpoint

For journalled quota mode, let checkpoint to flush dquot dirty data
and quota file data to guarntee persistence of all quota sysfile in
last checkpoint, by this way, we can avoid corrupting quota sysfile
when encountering SPO.

The implementation is as below:

1. add a global state SBI_QUOTA_NEED_FLUSH to indicate that there is
cached dquot metadata changes in quota subsystem, and later checkpoint
should:
 a) flush dquot metadata into quota file.
 b) flush quota file to storage to keep file usage be consistent.

2. add a global state SBI_QUOTA_NEED_REPAIR to indicate that quota
operation failed due to -EIO or -ENOSPC, so later,
 a) checkpoint will skip syncing dquot metadata.
 b) CP_QUOTA_NEED_FSCK_FLAG will be set in last cp pack to give a
    hint for fsck repairing.

3. add a global state SBI_QUOTA_SKIP_FLUSH, in checkpoint, if quota
data updating is very heavy, it may cause hungtask in block_operation().
To avoid this, if our retry time exceed threshold, let's just skip
flushing and retry in next checkpoint().

Signed-off-by: Weichao Guo <guoweichao@huawei.com>
Signed-off-by: Chao Yu <yuchao0@huawei.com>
[Jaegeuk Kim: avoid warnings and set fsck flag]
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>

1 files changed, 109 insertions, 12 deletions

diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c
index f7814bb26a13..af58b2cc21b8 100644
--- a/fs/f2fs/super.c
+++ b/fs/f2fs/super.c
@@ -1711,6 +1711,7 @@ repeat:
 				congestion_wait(BLK_RW_ASYNC, HZ/50);
 				goto repeat;
 			}
+			set_sbi_flag(F2FS_SB(sb), SBI_QUOTA_NEED_REPAIR);
 			return PTR_ERR(page);
 		}
 
@@ -1722,6 +1723,7 @@ repeat:
 		}
 		if (unlikely(!PageUptodate(page))) {
 			f2fs_put_page(page, 1);
+			set_sbi_flag(F2FS_SB(sb), SBI_QUOTA_NEED_REPAIR);
 			return -EIO;
 		}
 
@@ -1763,6 +1765,7 @@ retry:
 				congestion_wait(BLK_RW_ASYNC, HZ/50);
 				goto retry;
 			}
+			set_sbi_flag(F2FS_SB(sb), SBI_QUOTA_NEED_REPAIR);
 			break;
 		}
 
@@ -1799,6 +1802,12 @@ static qsize_t *f2fs_get_reserved_space(struct inode *inode)
 
 static int f2fs_quota_on_mount(struct f2fs_sb_info *sbi, int type)
 {
+	if (is_set_ckpt_flags(sbi, CP_QUOTA_NEED_FSCK_FLAG)) {
+		f2fs_msg(sbi->sb, KERN_ERR,
+			"quota sysfile may be corrupted, skip loading it");
+		return 0;
+	}
+
 	return dquot_quota_on_mount(sbi->sb, F2FS_OPTION(sbi).s_qf_names[type],
 					F2FS_OPTION(sbi).s_jquota_fmt, type);
 }
@@ -1869,7 +1878,14 @@ static int f2fs_enable_quotas(struct super_block *sb)
 		test_opt(F2FS_SB(sb), PRJQUOTA),
 	};
 
-	sb_dqopt(sb)->flags |= DQUOT_QUOTA_SYS_FILE | DQUOT_NOLIST_DIRTY;
+	if (is_set_ckpt_flags(F2FS_SB(sb), CP_QUOTA_NEED_FSCK_FLAG)) {
+		f2fs_msg(sb, KERN_ERR,
+			"quota file may be corrupted, skip loading it");
+		return 0;
+	}
+
+	sb_dqopt(sb)->flags |= DQUOT_QUOTA_SYS_FILE;
+
 	for (type = 0; type < MAXQUOTAS; type++) {
 		qf_inum = f2fs_qf_ino(sb, type);
 		if (qf_inum) {
@@ -1883,6 +1899,8 @@ static int f2fs_enable_quotas(struct super_block *sb)
 					"fsck to fix.", type, err);
 				for (type--; type >= 0; type--)
 					dquot_quota_off(sb, type);
+				set_sbi_flag(F2FS_SB(sb),
+						SBI_QUOTA_NEED_REPAIR);
 				return err;
 			}
 		}
@@ -1890,35 +1908,51 @@ static int f2fs_enable_quotas(struct super_block *sb)
 	return 0;
 }
 
-static int f2fs_quota_sync(struct super_block *sb, int type)
+int f2fs_quota_sync(struct super_block *sb, int type)
 {
+	struct f2fs_sb_info *sbi = F2FS_SB(sb);
 	struct quota_info *dqopt = sb_dqopt(sb);
 	int cnt;
 	int ret;
 
 	ret = dquot_writeback_dquots(sb, type);
 	if (ret)
-		return ret;
+		goto out;
 
 	/*
 	 * Now when everything is written we can discard the pagecache so
 	 * that userspace sees the changes.
 	 */
 	for (cnt = 0; cnt < MAXQUOTAS; cnt++) {
+		struct address_space *mapping;
+
 		if (type != -1 && cnt != type)
 			continue;
 		if (!sb_has_quota_active(sb, cnt))
 			continue;
 
-		ret = filemap_write_and_wait(dqopt->files[cnt]->i_mapping);
+		mapping = dqopt->files[cnt]->i_mapping;
+
+		ret = filemap_fdatawrite(mapping);
 		if (ret)
-			return ret;
+			goto out;
+
+		/* if we are using journalled quota */
+		if (is_journalled_quota(sbi))
+			continue;
+
+		ret = filemap_fdatawait(mapping);
+		if (ret)
+			set_sbi_flag(F2FS_SB(sb), SBI_QUOTA_NEED_REPAIR);
 
 		inode_lock(dqopt->files[cnt]);
 		truncate_inode_pages(&dqopt->files[cnt]->i_data, 0);
 		inode_unlock(dqopt->files[cnt]);
 	}
-	return 0;
+out:
+	if (ret)
+		set_sbi_flag(F2FS_SB(sb), SBI_QUOTA_NEED_REPAIR);
+	return ret;
 }
 
 static int f2fs_quota_on(struct super_block *sb, int type, int format_id,
@@ -1986,7 +2020,7 @@ void f2fs_quota_off_umount(struct super_block *sb)
 				"Fail to turn off disk quota "
 				"(type: %d, err: %d, ret:%d), Please "
 				"run fsck to fix it.", type, err, ret);
-			set_sbi_flag(F2FS_SB(sb), SBI_NEED_FSCK);
+			set_sbi_flag(F2FS_SB(sb), SBI_QUOTA_NEED_REPAIR);
 		}
 	}
 }
@@ -2003,6 +2037,61 @@ static void f2fs_truncate_quota_inode_pages(struct super_block *sb)
 	}
 }
 
+static int f2fs_dquot_commit(struct dquot *dquot)
+{
+	int ret;
+
+	ret = dquot_commit(dquot);
+	if (ret < 0)
+		set_sbi_flag(F2FS_SB(dquot->dq_sb), SBI_QUOTA_NEED_REPAIR);
+	return ret;
+}
+
+static int f2fs_dquot_acquire(struct dquot *dquot)
+{
+	int ret;
+
+	ret = dquot_acquire(dquot);
+	if (ret < 0)
+		set_sbi_flag(F2FS_SB(dquot->dq_sb), SBI_QUOTA_NEED_REPAIR);
+
+	return ret;
+}
+
+static int f2fs_dquot_release(struct dquot *dquot)
+{
+	int ret;
+
+	ret = dquot_release(dquot);
+	if (ret < 0)
+		set_sbi_flag(F2FS_SB(dquot->dq_sb), SBI_QUOTA_NEED_REPAIR);
+	return ret;
+}
+
+static int f2fs_dquot_mark_dquot_dirty(struct dquot *dquot)
+{
+	struct super_block *sb = dquot->dq_sb;
+	struct f2fs_sb_info *sbi = F2FS_SB(sb);
+	int ret;
+
+	ret = dquot_mark_dquot_dirty(dquot);
+
+	/* if we are using journalled quota */
+	if (is_journalled_quota(sbi))
+		set_sbi_flag(sbi, SBI_QUOTA_NEED_FLUSH);
+
+	return ret;
+}
+
+static int f2fs_dquot_commit_info(struct super_block *sb, int type)
+{
+	int ret;
+
+	ret = dquot_commit_info(sb, type);
+	if (ret < 0)
+		set_sbi_flag(F2FS_SB(sb), SBI_QUOTA_NEED_REPAIR);
+	return ret;
+}
 
 static int f2fs_get_projid(struct inode *inode, kprojid_t *projid)
 {
@@ -2012,11 +2101,11 @@ static int f2fs_get_projid(struct inode *inode, kprojid_t *projid)
 
 static const struct dquot_operations f2fs_quota_operations = {
 	.get_reserved_space = f2fs_get_reserved_space,
-	.write_dquot	= dquot_commit,
-	.acquire_dquot	= dquot_acquire,
-	.release_dquot	= dquot_release,
-	.mark_dirty	= dquot_mark_dquot_dirty,
-	.write_info	= dquot_commit_info,
+	.write_dquot	= f2fs_dquot_commit,
+	.acquire_dquot	= f2fs_dquot_acquire,
+	.release_dquot	= f2fs_dquot_release,
+	.mark_dirty	= f2fs_dquot_mark_dquot_dirty,
+	.write_info	= f2fs_dquot_commit_info,
 	.alloc_dquot	= dquot_alloc,
 	.destroy_dquot	= dquot_destroy,
 	.get_projid	= f2fs_get_projid,
@@ -2034,6 +2123,11 @@ static const struct quotactl_ops f2fs_quotactl_ops = {
 	.get_nextdqblk	= dquot_get_next_dqblk,
 };
 #else
+int f2fs_quota_sync(struct super_block *sb, int type)
+{
+	return 0;
+}
+
 void f2fs_quota_off_umount(struct super_block *sb)
 {
 }
@@ -3104,6 +3198,9 @@ try_onemore:
 		goto free_meta_inode;
 	}
 
+	if (__is_set_ckpt_flags(F2FS_CKPT(sbi), CP_QUOTA_NEED_FSCK_FLAG))
+		set_sbi_flag(sbi, SBI_QUOTA_NEED_REPAIR);
+
 	/* Initialize device list */
 	err = f2fs_scan_devices(sbi);
 	if (err) {