From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-ed1-f69.google.com (mail-ed1-f69.google.com [209.85.208.69]) by kanga.kvack.org (Postfix) with ESMTP id 7451F6B0270 for ; Wed, 10 Oct 2018 12:39:38 -0400 (EDT) Received: by mail-ed1-f69.google.com with SMTP id x44-v6so3488826edd.17 for ; Wed, 10 Oct 2018 09:39:38 -0700 (PDT) Received: from mx1.suse.de (mx2.suse.de. [195.135.220.15]) by mx.google.com with ESMTPS id r22-v6si6311036edm.167.2018.10.10.09.39.36 for (version=TLS1_2 cipher=ECDHE-RSA-AES128-GCM-SHA256 bits=128/128); Wed, 10 Oct 2018 09:39:36 -0700 (PDT) Subject: Re: [PATCH v5 2/2] mm/page_alloc: remove software prefetching in __free_pages_core References: <1538727006-5727-1-git-send-email-arunks@codeaurora.org> <1538727006-5727-2-git-send-email-arunks@codeaurora.org> From: Vlastimil Babka Message-ID: Date: Wed, 10 Oct 2018 18:36:52 +0200 MIME-Version: 1.0 In-Reply-To: <1538727006-5727-2-git-send-email-arunks@codeaurora.org> Content-Type: text/plain; charset=utf-8 Content-Language: en-US Content-Transfer-Encoding: 7bit Sender: owner-linux-mm@kvack.org List-ID: To: Arun KS , kys@microsoft.com, haiyangz@microsoft.com, sthemmin@microsoft.com, boris.ostrovsky@oracle.com, jgross@suse.com, akpm@linux-foundation.org, dan.j.williams@intel.com, mhocko@suse.com, iamjoonsoo.kim@lge.com, gregkh@linuxfoundation.org, osalvador@suse.de, malat@debian.org, kirill.shutemov@linux.intel.com, jrdr.linux@gmail.com, yasu.isimatu@gmail.com, mgorman@techsingularity.net, aaron.lu@intel.com, devel@linuxdriverproject.org, linux-kernel@vger.kernel.org, linux-mm@kvack.org, xen-devel@lists.xenproject.org Cc: vatsa@codeaurora.org, vinmenon@codeaurora.org, getarunks@gmail.com On 10/5/18 10:10 AM, Arun KS wrote: > They not only increase the code footprint, they actually make things > slower rather than faster. Remove them as contemporary hardware doesn't > need any hint. > > Suggested-by: Dan Williams > Signed-off-by: Arun KS Yeah, a tight loop with fixed stride is a trivial case for hw prefetcher. Acked-by: Vlastimil Babka > --- > mm/page_alloc.c | 6 +----- > 1 file changed, 1 insertion(+), 5 deletions(-) > > diff --git a/mm/page_alloc.c b/mm/page_alloc.c > index 7ab5274..90db431 100644 > --- a/mm/page_alloc.c > +++ b/mm/page_alloc.c > @@ -1258,14 +1258,10 @@ void __free_pages_core(struct page *page, unsigned int order) > struct page *p = page; > unsigned int loop; > > - prefetchw(p); > - for (loop = 0; loop < (nr_pages - 1); loop++, p++) { > - prefetchw(p + 1); > + for (loop = 0; loop < nr_pages ; loop++, p++) { > __ClearPageReserved(p); > set_page_count(p, 0); > } > - __ClearPageReserved(p); > - set_page_count(p, 0); > > page_zone(page)->managed_pages += nr_pages; > set_page_refcounted(page); >