* Theodore Y. Ts'o <tytso@xxxxxxx>:
> On Thu, Jul 26, 2018 at 01:16:17PM -0400, Eric Whitney wrote:
> > As mentioned in today's ext4 concall, I've been seeing generic/231 failures
> > when running kvm-xfstests on recent 4.18 kernels, beginning with rc4 and
> > continuing through rc5 and rc6. To get the failure, it's necessary to
> > run generic/230 and 231 in sequence. Running generic/231 alone and
> > repeatedly does not reproduce the failure for me.
> >
> > The failure bisects to a patch in 4.18-rc4:
> > (8844618d8aa7) ext4: only look at the bg_flags field if it is valid
>
> Thanks, I found the problem with that patch. Patch follows...
>
> commit e21d00b2ed3f480224b3b2a620326cc5ac1ad5f2
> Author: Theodore Ts'o <tytso@xxxxxxx>
> Date: Sun Jul 28 23:12:04 2018 -0400
>
> ext4: fix check to prevent initializing reserved inodes
>
> Commit 8844618d8aa7: "ext4: only look at the bg_flags field if it is
> valid" will complain if block group zero does not have the
> EXT4_BG_INODE_ZEROED flag set. Unfortunately, this is not correct,
> since a freshly created file system has this flag cleared. It gets
> almost immediately after the file system is mounted read-write --- but
> the following somewhat unlikely sequence will end up triggering a
> false positive report of a corrupted file system:
>
> mkfs.ext4 /dev/vdc
> mount -o ro /dev/vdc /vdc
> mount -o remount,rw /dev/vdc
>
> Instead, when initializing the inode table for block group zero, test
> to make sure that itable_unused count is not too large, since that is
> the case that will result in some or all of the reserved inodes
> getting cleared.
>
> This fixes the failures reported by Eric Whiteney when running
> generic/230 and generic/231 in the the nojournal test case.
>
> Fixes: 8844618d8aa7 ("ext4: only look at the bg_flags field if it is valid")
> Reported-by: Eric Whitney <enwlinux@xxxxxxxxx>
> Signed-off-by: Theodore Ts'o <tytso@xxxxxxx>
>
> diff --git a/fs/ext4/ialloc.c b/fs/ext4/ialloc.c
> index 9429242cf4f7..dffd21de2694 100644
> --- a/fs/ext4/ialloc.c
> +++ b/fs/ext4/ialloc.c
> @@ -1388,7 +1388,10 @@ int ext4_init_inode_table(struct super_block *sb, ext4_group_t group,
> ext4_itable_unused_count(sb, gdp)),
> sbi->s_inodes_per_block);
>
> - if ((used_blks < 0) || (used_blks > sbi->s_itb_per_group)) {
> + if ((used_blks < 0) || (used_blks > sbi->s_itb_per_group) ||
> + ((group == 0) && ((EXT4_INODES_PER_GROUP(sb) -
> + ext4_itable_unused_count(sb, gdp)) <
> + EXT4_FIRST_INO(sb)))) {
> ext4_error(sb, "Something is wrong with group %u: "
> "used itable blocks: %d; "
> "itable unused count: %u",
> diff --git a/fs/ext4/super.c b/fs/ext4/super.c
> index 92de019c117d..d4a218ba626c 100644
> --- a/fs/ext4/super.c
> +++ b/fs/ext4/super.c
> @@ -3162,14 +3162,8 @@ static ext4_group_t ext4_has_uninit_itable(struct super_block *sb)
> if (!gdp)
> continue;
>
> - if (gdp->bg_flags & cpu_to_le16(EXT4_BG_INODE_ZEROED))
> - continue;
> - if (group != 0)
> + if (!(gdp->bg_flags & cpu_to_le16(EXT4_BG_INODE_ZEROED)))
> break;
> - ext4_error(sb, "Inode table for bg 0 marked as "
> - "needing zeroing");
> - if (sb_rdonly(sb))
> - return ngroups;
> }
>
> return group;
After applying that patch to 4.18-rc6, I'm no longer seeing failures with
kvm-xfstests in the nojournal test case. Looks good.
Tested-by: Eric Whitney <enwlinux@xxxxxxxxx>
Thanks!
Eric
