Re: [PATCH V3 01/27] kfifo: add kfifo_alloc_node() helper for NUMA awareness

[Andrew Morton]

On Thu, 13 Nov 2025 09:21:41 +0800 Ming Lei <ming.lei@xxxxxxxxxx> wrote:

> > > +#define kfifo_alloc_node(fifo, size, gfp_mask, node) \
> > > +__kfifo_int_must_check_helper( \
> > > +({ \
> > > +	typeof((fifo) + 1) __tmp = (fifo); \
> > > +	struct __kfifo *__kfifo = &__tmp->kfifo; \
> > > +	__is_kfifo_ptr(__tmp) ? \
> > > +	__kfifo_alloc_node(__kfifo, size, sizeof(*__tmp->type), gfp_mask, node) : \
> > > +	-EINVAL; \
> > > +}) \
> > > +)
> > 
> > Well this is an eyesore.  Do we really need it?  It seems to be here so
> > we can check for a programming bug?  Well, don't add programming bugs!
> > 
> > I'm actually not enjoying the existence of __is_kfifo_ptr() at all. 
> > What  is it all doing?  It's a FIFO for heck's sake, why is this so hard.
> 
> It is basically a clone of existing kfifo_alloc().
> 
> Do we need to clean kfifo_alloc() first? Otherwise I'd keep the same
> pattern with existing definitions.

Is OK, I was just whining.  Possibly de-over-engineering kifo is a
separate project.