Re: [PATCH 01/10] Add parse_integer() (replacement for simple_strto*())

[Rasmus Villemoes <linux@rasmusvillemoes.dk>]

On Mon, May 04 2015, Alexey Dobriyan <adobriyan@gmail.com> wrote:

>> There are lots of callers of memparse(), and I don't think any of them
>> are prepared to handle *endp ending up pointing before the passed-in
>> string (-EINVAL == -22, -ERANGE == -34). I can easily see how that could
>> lead to an infinite loop, maybe worse.
>
> Yeah, possible bug could become worse, I'll add error checking,
> but, seriously, you're defending this :^)
>
> 	case Opt_nr_inodes:
> ===>		/* memparse() will accept a K/M/G without a digit */
> ===>		if (!isdigit(*args[0].from))
> ===>			goto bad_val;
> 		pconfig->nr_inodes = memparse(args[0].from, &rest);
> 		break;
>

No, I'm not defending memparse(), simple_strto* or any of their
callers. I'm just trying to say that you can't change the semantics of
memparse() without considering all its callers.

I don't think there's any way to "add error checking" and preserve the
exact memparse() semantic - in other words, I don't think simple_strto*
can actually be implemented in terms of parse_integer. But that's not a
bad thing - we want to get rid of those.

> memparse() is misdesigned in the same sense strtoul() is misdesigned.
> Every "memparse(s, NULL)" user is a bug for example.

Yes, memparse is misdesigned, since it doesn't have a way to indicate
error. That leads me to: There's no point in adding error checking to
the integer parsing part without also checking the left shifts for
overflow.

I think the right approach is to rename memparse to legacy_memparse and
introduce a memparse with semantics that allow error checking. One could
start by introducing that under the name sane_memparse. But there are
probably lots of simple_strto*() uses that are easier to eliminate.

Rasmus