Re: SLUB: Return ZERO_SIZE_PTR for kmalloc(0)

[Cyrill Gorcunov <gorcunov@gmail.com>]

[Pekka Enberg - Mon, Jun 04, 2007 at 07:37:01PM +0300]
| On 6/4/07, Linus Torvalds <torvalds@linux-foundation.org> wrote:
| >Well, the red-zones won't catch readers, and more importantly, even for
| >writers they are *really* inconvenient, because it will just tell you
| >something bad happened, it won't tell you *where* it happened.
| 
| True.
| 
| On 6/4/07, Linus Torvalds <torvalds@linux-foundation.org> wrote:
| >Since comparing the addresses of two zero-sized allocations is insane and
| >not done _anyway_, it's just much better to return an invalid address.
| 
| Then we might as well return your regular NULL pointer for zero-length
| allocations as you can't do anything sane with ZERO_SIZE_PTR either.

Hi Pekka,

you know, I'm absolutely agree with you. Hey, kernel hackers don't blame
me ;). But lets just take an example from a real life: I'm asking for
someone to give me 50 cents he would probably give me this. But if I'm
asking someone to give me 0 cents - he''ll give me nothing!!! Nobody
giving me a spec. papper on which "zero cents" is written. :)
So the code

	p = kmalloc(0);
	if (!p) {
		return -ENOMEM;
	}

is absolutely right - we physically have _no_ zero-sized memory. And as result
we have no reason to keep spec. slab to track zero-sized allocs.

Please, don't get me wrong, I'm not a kernel specialist...
And sorry for meddling into coversation.

		Cyrill