Re: [PATCH v1] fs/super.c: Add NULL check for type in iterate_supers_type

[Al Viro]

On Thu, Apr 03, 2025 at 11:10:02AM +0800, Xiaole He wrote:
> Thank you for your thoughtfully feedback.
> I think you are right, and I'm sorry for my pedantic anxiety.
> So if we just ignore this patch for now, or I should submit a bug
> report to kernel community in order to invite more thorough fix?
> Thanks for your patient again.

I don't believe that adding random checks would make any sense -
same as for any library function, really.

Having the documentation slightly more clear would make sense,
though; currently it's

 *      iterate_supers_type - call function for superblocks of given type
 *      @type: fs type
 *      @f: function to call
 *      @arg: argument to pass to it
 *
 *      Scans the superblock list and calls given function, passing it
 *      locked superblock and given argument.

and description could've been better.  The weakest part in there is,
IMO, "the superblock list" - there is a global list of all superblocks
(inventively called 'super_blocks'), but that's not what gets scanned;
the list of superblocks of given type (type->fs_supers) we iterate
through.

Something along the lines of "Call given callback @f for all superblocks
of given type.  The first argument passed to @f points to a locked superblock
that belongs to @type; the second is a caller-supplied opaque pointer @arg.
The caller is responsible for passing @arg that would make a valid second
argument for @f - compiler can't help here" might be a starting point,
but I'm not up to turning that into proper English.