On Tue, Aug 02, 2022 at 10:10:29AM +0800, Michael Wu wrote: > The following error occurs when mounting the ext4 image made by image > making tool: > "ext4_init_inode_table:1301:comm ext4lazyinit:Something is wrong with group > 0: used itable blocks: 491; itable unused count: 0." > > Currently all the inodes in block group0 and ext4 image is divided by > s_inodes_per_group. That leads to a hazard: we can't ensure all > s_inodes_per_group are divisible by s_inodes_per_block. For example, when > the s_inodes_per_group (equals to 7851) is divided by s_inodes_per_block > (which is 16), because 7851 is undivisible by 16, we get the wrong result > 490, while 491 is expected. > > So, we suggest that s_itb_per_group should equal to > DIV_ROUND_UP(s_inodes_per_group, s_inodes_per_block) instead of directly > getting the result from s_inodes_per_group/s_inodes_per_block. Hi Michael, mke2fs is making sure that we completely fill the inote table blocks. This is a corrupted image and so AFAICT ext4 is doing the right thing here. There does not seem to be a problem to fix, unless you can somehow trick mke2fs to make a file system like this. -Lukas > > Signed-off-by: Michael Wu <michael@xxxxxxxxxxxxxxxxx> > --- > fs/ext4/super.c | 4 ++-- > 1 file changed, 2 insertions(+), 2 deletions(-) > > diff --git a/fs/ext4/super.c b/fs/ext4/super.c > index 845f2f8aee5f..76cbd638ea10 100644 > --- a/fs/ext4/super.c > +++ b/fs/ext4/super.c > @@ -4796,8 +4796,8 @@ static int __ext4_fill_super(struct fs_context *fc, struct super_block *sb) > sbi->s_inodes_per_group); > goto failed_mount; > } > - sbi->s_itb_per_group = sbi->s_inodes_per_group / > - sbi->s_inodes_per_block; > + sbi->s_itb_per_group = DIV_ROUND_UP(sbi->s_inodes_per_group, > + sbi->s_inodes_per_block); > sbi->s_desc_per_block = blocksize / EXT4_DESC_SIZE(sb); > sbi->s_sbh = bh; > sbi->s_mount_state = le16_to_cpu(es->s_state) & ~EXT4_FC_REPLAY; > -- > 2.29.0 >
