On Tue 17-05-22 19:40:11, Theodore Ts'o wrote: > On Thu, Apr 28, 2022 at 08:31:37PM +0200, Jan Kara wrote: > > Before splitting a directory block verify its directory entries are sane > > so that the splitting code does not access memory it should not. > > This commit fails to build due to an undefined variable. It's fixed > with this hunk in the next patch, which needs to be brought back into > this commit: > > diff --git a/fs/ext4/namei.c b/fs/ext4/namei.c > index 5951e9bb348e..7286472e9558 100644 > --- a/fs/ext4/namei.c > +++ b/fs/ext4/namei.c > @@ -1278,7 +1278,7 @@ static int dx_make_map(struct inode *dir, struct buffer_head *bh, > count++; > cond_resched(); > } > - de = ext4_next_entry(de, blocksize); > + de = ext4_next_entry(de, dir->i_sb->s_blocksize); > } > return count; > } > > I was thinking about folding in this change and apply the patch with > that change --- and I may yet do that --- but it looks like there's a > bigger problem with this patch series, which is that it's causing a > crash when running ext4/052 due to what appears to be a smashed stack. > More about that in the reply to patch 2/2 of this series.... Yup, I'll fix that. Thanks for catching this. Honza -- Jan Kara <jack@xxxxxxxx> SUSE Labs, CR