Re: [RFC RFT PATCH] ocfs2: Mark inode bad upon validation failure during read

[Heming Zhao]

On Fri, Oct 31, 2025 at 09:54:41AM -0400, Albin Babu Varghese wrote:
> Hi Heming,
> 
> > I support adding make_bad_inode() in ocfs2_read_inode_block_full().
> > ocfs2_read_locked_inode() calls ocfs2_read_inode_block[_full] to read the inode
> > from disk. However, ocfs2_read_inode_block[_full] have many callers, and in
> > current code, only ocfs2_read_locked_inode() marks the inode as bad. All others
> > forget to set the bad_inode.
> > 
> > The 'forbid' write operations when read-only mode is worth another patch, and
> > I plan to create this patch. This patch adds a similar ext4_emergency_state()
> > function for ocfs2.
> 
> We're working on this as part of the Linux Kernel Mentorship Program, and we'd
> love to take on implementing the read-only check if it's not overly
> complicated. We're just beginners, but we thought it would be a great learning
> experience to work on this following the ext4 pattern you mentioned - if you
> haven't already started working on it by the time you see this reply.

I haven't started the patch job, you are welcome to take it.

- Heming
> 
> > Therefore, your original patch looks good to me. I will provide my Reivewed-by
> > for it.
> 
> Thank you so much for your review and the Reviewed-by tag
> 
> Cheers,
> 	Albin