From mboxrd@z Thu Jan  1 00:00:00 1970
X-Msuck: nntp://news.gmane.org/gmane.linux.lib.musl.general/7
Path: news.gmane.org!not-for-mail
From: Rich Felker <dalias@aerifal.cx>
Newsgroups: gmane.linux.lib.musl.general
Subject: Re: linux/sbrk is wrong
Date: Sun, 22 May 2011 12:09:14 -0400
Message-ID: <20110522160914.GR277@brightrain.aerifal.cx>
References: <20110522145555.GE6142@port70.net>
Reply-To: musl@lists.openwall.com
NNTP-Posting-Host: lo.gmane.org
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-Trace: dough.gmane.org 1306081160 7280 80.91.229.12 (22 May 2011 16:19:20 GMT)
X-Complaints-To: usenet@dough.gmane.org
NNTP-Posting-Date: Sun, 22 May 2011 16:19:20 +0000 (UTC)
To: musl@lists.openwall.com
Original-X-From: musl-return-90-gllmg-musl=m.gmane.org@lists.openwall.com Sun May 22 18:19:17 2011
Return-path: <musl-return-90-gllmg-musl=m.gmane.org@lists.openwall.com>
Envelope-to: gllmg-musl@lo.gmane.org
Original-Received: from mother.openwall.net ([195.42.179.200])
	by lo.gmane.org with smtp (Exim 4.69)
	(envelope-from <musl-return-90-gllmg-musl=m.gmane.org@lists.openwall.com>)
	id 1QOBN6-0003y1-Vc
	for gllmg-musl@lo.gmane.org; Sun, 22 May 2011 18:19:17 +0200
Original-Received: (qmail 24228 invoked by uid 550); 22 May 2011 16:19:16 -0000
Mailing-List: contact musl-help@lists.openwall.com; run by ezmlm
Precedence: bulk
List-Post: <mailto:musl@lists.openwall.com>
List-Help: <mailto:musl-help@lists.openwall.com>
List-Unsubscribe: <mailto:musl-unsubscribe@lists.openwall.com>
List-Subscribe: <mailto:musl-subscribe@lists.openwall.com>
Original-Received: (qmail 24215 invoked from network); 22 May 2011 16:19:16 -0000
Content-Disposition: inline
In-Reply-To: <20110522145555.GE6142@port70.net>
User-Agent: Mutt/1.5.21 (2010-09-15)
Xref: news.gmane.org gmane.linux.lib.musl.general:7
Archived-At: <http://permalink.gmane.org/gmane.linux.lib.musl.general/7>

On Sun, May 22, 2011 at 04:55:55PM +0200, Szabolcs Nagy wrote:
> linux brk syscall returns the end pointer
> so if it fails then the old one
> if it succeeds then the new one
> 
> the current implementation of sbrk returns
> the end of the allocated area instead of
> the begining

IMO there's no way for correct code to use these functions anyway, but
I'll fix them anyway.

>  int brk(void *end)
>  {
> -	return -(syscall(SYS_brk, end) == -1);
> +	return -((void *)syscall(SYS_brk, end) != end);
>  }

This looks fine.

>  void *sbrk(ptrdiff_t inc)
>  {
> -	return (void *)syscall(SYS_brk, syscall(SYS_brk, 0)+inc);
> +	void *p = (void *)syscall(SYS_brk, 0);
> +
> +	if (inc == 0)
> +		return p;
> +	if ((void *)syscall(SYS_brk, p+inc) == p+inc)

This line is invalid C. p+inc is not defined since p has type void *.
The most correct would probably be to use uintptr_t internally
everywhere and just cast to void * in the return statement.

Rich