Re: [PATCH v2 0/2] Lock and Pointer guards

[Peter Zijlstra]

On Thu, Jun 08, 2023 at 07:25:27PM -0700, Linus Torvalds wrote:
> On Thu, Jun 8, 2023 at 1:06 PM Peter Zijlstra <peterz@xxxxxxxxxxxxx> wrote:
> >
> >         struct obj *p __free(kfree) = kmalloc(...);
> 
> Yeah, the above actually looks really good to me - I like the naming
> here, and the use looks very logical to me.
> 
> Of course, maybe once I see the patches that use this I go "uhh", but
> at least for now I think you've hit on a rather legible syntax.
> 
> I'm still a bit unsure of the "no_free_ptr(p)" naming, but at least
> it's pretty clear about what it does.
> 
> So my only worry is that it's not pretty and to the point like your
> "__free(kfree)" syntax.
> 
> But it feels workable and not misleading, so unless somebody can come
> up with a better name, I think it's ok.

#define return_ptr(p) \
	return no_free_ptr(p)


	struct obj *p __free(kfree) = kmalloc(...);
	if (!p)
		return ERR_PTR(-ENOMEM);

	...

	return_ptr(p);

?