From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from mailman by lists.gnu.org with tmda-scanned (Exim 4.43)
	id 1MbX1j-0004Q1-Vx
	for qemu-devel@nongnu.org; Thu, 13 Aug 2009 05:55:20 -0400
Received: from exim by lists.gnu.org with spam-scanned (Exim 4.43)
	id 1MbX1i-0004PM-6U
	for qemu-devel@nongnu.org; Thu, 13 Aug 2009 05:55:19 -0400
Received: from [199.232.76.173] (port=54429 helo=monty-python.gnu.org)
	by lists.gnu.org with esmtp (Exim 4.43) id 1MbX1i-0004PI-0c
	for qemu-devel@nongnu.org; Thu, 13 Aug 2009 05:55:18 -0400
Received: from [84.20.150.76] (port=59989 helo=naru.obs2.net)
	by monty-python.gnu.org with esmtps
	(TLS-1.0:DHE_RSA_AES_256_CBC_SHA1:32) (Exim 4.60)
	(envelope-from <nchip@kos.to>) id 1MbX1h-0001Oi-9y
	for qemu-devel@nongnu.org; Thu, 13 Aug 2009 05:55:17 -0400
Date: Thu, 13 Aug 2009 12:55:12 +0300
From: Riku Voipio <riku.voipio@iki.fi>
Subject: Re: [Qemu-devel] [PATCH] linux-user: elfload: allocate and
	initialize memsz-filesz gap
Message-ID: <20090813095512.GA6866@kos.to>
References: <200907121502.39025.jcmvbkbc@gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <200907121502.39025.jcmvbkbc@gmail.com>
List-Id: qemu-devel.nongnu.org
List-Unsubscribe: <http://lists.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=unsubscribe>
List-Archive: <http://lists.gnu.org/pipermail/qemu-devel>
List-Post: <mailto:qemu-devel@nongnu.org>
List-Help: <mailto:qemu-devel-request@nongnu.org?subject=help>
List-Subscribe: <http://lists.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=subscribe>
To: Max Filippov <jcmvbkbc@gmail.com>
Cc: qemu-devel@nongnu.org

On Sun, Jul 12, 2009 at 03:02:38PM +0400, Max Filippov wrote:
> I'm having ppc ELF binaries that I'm running on x86 in linux-user emulation.
> Some of these binaries work fine, but others segfault.
> The latter binaries have segments with p_filesz < p_memsz and sections like .bss in this gap.
> Segfaults usually happen in attempt to access address within this gap. 
> According to the linux-user/elfload.c only first p_filesz bytes of such segments are mmaped and mprotected.

> This patch mmaps p_memsz bytes and then zero out last p_memsz - p_filesz bytes of such segments.
> Is there a better way to do it?

After applying this patch, any arm-linux binaries on x86 fail run.

> Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
> ---
> diff -burN qemu-snapshot-2009-07-11_r7249-orig/linux-user/elfload.c qemu-snapshot-2009-07-11_r7249/linux-user/elfload.c
> --- qemu-snapshot-2009-07-11_r7249-orig/linux-user/elfload.c	2009-07-11 06:12:23.000000000 +0400
> +++ qemu-snapshot-2009-07-11_r7249/linux-user/elfload.c	2009-07-12 04:47:47.000000000 +0400
> @@ -1380,7 +1380,7 @@
>          }
>  
>          error = target_mmap(TARGET_ELF_PAGESTART(load_bias + elf_ppnt->p_vaddr),
> -                            (elf_ppnt->p_filesz +
> +                            (elf_ppnt->p_memsz +
>                               TARGET_ELF_PAGEOFFSET(elf_ppnt->p_vaddr)),
>                              elf_prot,
>                              (MAP_FIXED | MAP_PRIVATE | MAP_DENYWRITE),
> @@ -1392,6 +1392,20 @@
>              exit(-1);
>          }
>  
> +        if(elf_ppnt->p_memsz > elf_ppnt->p_filesz) {
> +            abi_ulong pg = TARGET_ELF_PAGESTART(load_bias + elf_ppnt->p_vaddr) +
> +                (elf_ppnt->p_filesz +
> +                 TARGET_ELF_PAGEOFFSET(elf_ppnt->p_vaddr));
> +            abi_ulong sz = elf_ppnt->p_memsz - elf_ppnt->p_filesz;
> +
> +            void *p = lock_user(PAGE_READ | PAGE_WRITE, pg, sz, 0);
> +
> +            if (p) {
> +                memset(p, 0, sz);
> +                unlock_user(p, pg, sz);
> +            }
> +        }
> +
>  #ifdef LOW_ELF_STACK
>          if (TARGET_ELF_PAGESTART(elf_ppnt->p_vaddr) < elf_stack)
>              elf_stack = TARGET_ELF_PAGESTART(elf_ppnt->p_vaddr);
> 
> 
> Thanks.
> -- Max
>