Skip to contentSkip to conversation

ext4: fix online resize's handling of a too-small final block group

PATCH New Theodore Ts'o 1 participant 2 messages Sep 4, 2018

Subsystems (2)
lore
Share & tools
lore

Theodore Ts'o

author maintainer
#0 lore raw

PATCH 1/2 ext4: fix online resize's handling of a too-small final block group

Avoid growing the file system to an extent so that the last block group is too small to hold all of the metadata that must be stored in the block group.

This problem can be triggered with the following reproducer:

umount /mnt

mke2fs -F -m0 -b 4096 -t ext4 -O resize_inode,^has_journal \
	-E resize=1073741824 /tmp/foo.img 128M

mount /tmp/foo.img /mnt
truncate --size 1708M /tmp/foo.img
resize2fs /dev/loop0 295400
umount /mnt
e2fsck -fy /tmp/foo.img

---
 fs/ext4/resize.c | 20 ++++++++++++++++++++
 1 file changed, 20 insertions(+)
fs/ext4/resize.c +20 -0
index e5fb38451a73..33655a6eff4d 100644
@@ -1986,6 +1986,26 @@ int ext4_resize_fs(struct super_block *sb, ext4_fsblk_t n_blocks_count)
 		}
 	}
 
+	/*
+	 * Make sure the last group has enough space so that it's
+	 * guaranteed to have enough space for all metadata blocks
+	 * that it might need to hold.  (We might not need to store
+	 * the inode table blocks in the last block group, but there
+	 * will be cases where this might be needed.)
+	 */
+	if ((ext4_group_first_block_no(sb, n_group) +
+	     ext4_group_overhead_blocks(sb, n_group) + 2 +
+	     sbi->s_itb_per_group + sbi->s_cluster_ratio) >= n_blocks_count) {
+		n_blocks_count = ext4_group_first_block_no(sb, n_group);
+		n_group--;
+		n_blocks_count_retry = 0;
+		if (resize_inode) {
+			iput(resize_inode);
+			resize_inode = NULL;
+		}
+		goto retry;
+	}
+
 	/* extend the last group */
 	if (n_group == o_group)
 		add = n_blocks_count - o_blocks_count;
Signature · 1 line
2.18.0.rc0
author maintainer
#1 lore raw

PATCH 2/2 ext4: fix online resizing for bigalloc file systems with a 1k block size

An online resize of a file system with the bigalloc feature enabled and a 1k block size would be refused since ext4_resize_begin() did not understand s_first_data_block is 0 for all bigalloc file systems, even when the block size is 1k.

---
 fs/ext4/resize.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
fs/ext4/resize.c +2 -1
index 33655a6eff4d..ebbc663d0798 100644
@@ -19,6 +19,7 @@
 
 int ext4_resize_begin(struct super_block *sb)
 {
+	struct ext4_sb_info *sbi = EXT4_SB(sb);
 	int ret = 0;
 
 	if (!capable(CAP_SYS_RESOURCE))
@@ -29,7 +30,7 @@ int ext4_resize_begin(struct super_block *sb)
          * because the user tools have no way of handling this.  Probably a
          * bad time to do it anyways.
          */
-	if (EXT4_SB(sb)->s_sbh->b_blocknr !=
+	if (EXT4_B2C(sbi, sbi->s_sbh->b_blocknr) !=
 	    le32_to_cpu(EXT4_SB(sb)->s_es->s_first_data_block)) {
 		ext4_warning(sb, "won't resize using backup superblock at %llu",
 			(unsigned long long)EXT4_SB(sb)->s_sbh->b_blocknr);
Signature · 1 line
2.18.0.rc0