Re: [PATCH] statmount: add flag to retrieve unescaped options

[Christian Brauner <brauner@kernel.org>]

On Thu, Nov 14, 2024 at 03:53:44PM +0100, Miklos Szeredi wrote:
> On Wed, 13 Nov 2024 at 17:31, Christian Brauner <brauner@kernel.org> wrote:
> 
> > Please take a look at the top of #vfs.misc and tell me whether this is
> > ok with you.
> 
> One more thing I'd do is:
> 
> -       if (seq->count == start)
> +       if (seq->count <= start + 1)
> 
> This would handle the (broken) case of just a single comma in the
> options.  So it's not a bug fix per-se, but would make it clear that
> the loop will run at least once, and so the seq->count calculation at
> the end won't be skewed.
> 
> The buf_start calculation could also be moved further down before the
> loop, where it's actually used.
> 
> I don't find the variable naming much better, how about taking the
> naming from string_unescape:
> 
> opt_start -> src - pointer to escaped string
> buf_start -> dst - pointer to de-escaped string
> 
> The others make sense.

Fine by me. I'm about to send pull-requests so I would suggest you send
a cleanup that I can include post-rc1 so we don't have to rebase right
now.