From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([2001:4830:134:3::10]:47758)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <pbonzini@redhat.com>) id 1bjqek-0005oH-Fj
	for qemu-devel@nongnu.org; Tue, 13 Sep 2016 12:33:59 -0400
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <pbonzini@redhat.com>) id 1bjqef-0007zS-VC
	for qemu-devel@nongnu.org; Tue, 13 Sep 2016 12:33:58 -0400
Received: from mx1.redhat.com ([209.132.183.28]:58124)
	by eggs.gnu.org with esmtp (Exim 4.71)
	(envelope-from <pbonzini@redhat.com>) id 1bjqef-0007yb-QU
	for qemu-devel@nongnu.org; Tue, 13 Sep 2016 12:33:53 -0400
References: <1473783005-113609-1-git-send-email-pbonzini@redhat.com>
	<1473783005-113609-11-git-send-email-pbonzini@redhat.com>
	<ac536bd0-96db-b023-d0b4-4694a4eea618@twiddle.net>
From: Paolo Bonzini <pbonzini@redhat.com>
Message-ID: <2734bd7f-48d4-a20c-ab8c-16b56bb370ed@redhat.com>
Date: Tue, 13 Sep 2016 18:33:50 +0200
MIME-Version: 1.0
In-Reply-To: <ac536bd0-96db-b023-d0b4-4694a4eea618@twiddle.net>
Content-Type: text/plain; charset=windows-1252
Content-Transfer-Encoding: 7bit
Subject: Re: [Qemu-devel] [PATCH 10/10] cutils: Rewrite x86 buffer zero
 checking
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: Richard Henderson <rth@twiddle.net>, qemu-devel@nongnu.org



On 13/09/2016 18:27, Richard Henderson wrote:
> On 09/13/2016 09:10 AM, Paolo Bonzini wrote:
>> @@ -177,16 +231,15 @@ bool test_buffer_is_zero_next_accel(void)
>>  
>>  static bool select_accel_fn(const void *buf, size_t len)
>>  {
>> -    uintptr_t ibuf = (uintptr_t)buf;
>>  #ifdef CONFIG_AVX2_OPT
>> -    if (len % 128 == 0 && ibuf % 32 == 0 && (cpuid_cache & CACHE_AVX2)) {
>> +    if (len >= 128 && (cpuid_cache & CACHE_AVX2)) {
>>          return buffer_zero_avx2(buf, len);
>>      }
>> -    if (len % 64 == 0 && ibuf % 16 == 0 && (cpuid_cache & CACHE_SSE4)) {
>> +    if (len >= 64 && (cpuid_cache & CACHE_SSE4)) {
>>          return buffer_zero_sse4(buf, len);
>>      }
>>  #endif
>> -    if (len % 64 == 0 && ibuf % 16 == 0 && (cpuid_cache & CACHE_SSE2)) {
>> +    if (len >= 64 && (cpuid_cache & CACHE_SSE2)) {
>>          return buffer_zero_sse2(buf, len);
>>      }
> 
> You've dropped a major change to select_accel_fn here.
> 
> (1) The avx2 routine, as written, can support len >= 64, therefore a common
> test works for all of the vectorized functions.
> 
> (2) I had saved the pointer to the routine, so that we didn't have to
> repeatedly test multiple cpuid_cache bits.

Can you send a replacement for this patch only?

Thanks,

Paolo