From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([2001:4830:134:3::10]:52275)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <rth7680@gmail.com>) id 1Zwt6A-0001fQ-JJ
	for qemu-devel@nongnu.org; Thu, 12 Nov 2015 09:43:39 -0500
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <rth7680@gmail.com>) id 1Zwt66-0001TM-FH
	for qemu-devel@nongnu.org; Thu, 12 Nov 2015 09:43:38 -0500
Received: from mail-wm0-x235.google.com ([2a00:1450:400c:c09::235]:35079)
	by eggs.gnu.org with esmtp (Exim 4.71)
	(envelope-from <rth7680@gmail.com>) id 1Zwt66-0001Su-8k
	for qemu-devel@nongnu.org; Thu, 12 Nov 2015 09:43:34 -0500
Received: by wmdw130 with SMTP id w130so157204698wmd.0
	for <qemu-devel@nongnu.org>; Thu, 12 Nov 2015 06:43:33 -0800 (PST)
Sender: Richard Henderson <rth7680@gmail.com>
References: <1447123907-26750-1-git-send-email-liang.z.li@intel.com>
	<1447123907-26750-2-git-send-email-liang.z.li@intel.com>
From: Richard Henderson <rth@twiddle.net>
Message-ID: <5644A591.90605@twiddle.net>
Date: Thu, 12 Nov 2015 15:43:29 +0100
MIME-Version: 1.0
In-Reply-To: <1447123907-26750-2-git-send-email-liang.z.li@intel.com>
Content-Type: text/plain; charset=windows-1252; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [Qemu-devel] [v2 1/2] cutils: add avx2 instruction optimization
List-Id: <qemu-devel.nongnu.org>
List-Unsubscribe: <https://lists.nongnu.org/mailman/options/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=unsubscribe>
List-Archive: <http://lists.nongnu.org/archive/html/qemu-devel>
List-Post: <mailto:qemu-devel@nongnu.org>
List-Help: <mailto:qemu-devel-request@nongnu.org?subject=help>
List-Subscribe: <https://lists.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=subscribe>
To: Liang Li <liang.z.li@intel.com>, qemu-devel@nongnu.org
Cc: amit.shah@redhat.com, pbonzini@redhat.com, mst@redhat.com, quintela@redhat.com

On 11/10/2015 03:51 AM, Liang Li wrote:
> +__asm__(".type can_use_buffer_find_nonzero_offset, \%gnu_indirect_function");
> +__asm__(".type buffer_find_nonzero_offset, \%gnu_indirect_function");
> +
> +
> +void *can_use_buffer_find_nonzero_offset_ifunc(void) \
> +                     __asm__("can_use_buffer_find_nonzero_offset");
> +
> +void *buffer_find_nonzero_offset_ifunc(void) \
> +                     __asm__("buffer_find_nonzero_offset");


Not keen on this.  You can use the ifunc attribute instead of inline asm, and 
the target attribute to enable per-function use of avx2.  And if neither are 
supported, due to compiler limitations, I don't think you should attempt to 
work around that.


r~