From mboxrd@z Thu Jan 1 00:00:00 1970 From: bill4carson@gmail.com (bill4carson) Date: Wed, 01 Feb 2012 09:56:50 +0800 Subject: [RFC] ARM hugetlb support In-Reply-To: References: <1327910238-18704-1-git-send-email-bill4carson@gmail.com> Message-ID: <4F289BE2.9020404@gmail.com> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On 2012?01?31? 17:29, Catalin Marinas wrote: > Hi, > > On 30 January 2012 07:57, wrote: >> This patch aims to support huge page for ARM. For now, 2MB(two 1MB page)/16MB >> huge page are supported, Versatile Express Cortex-A9x4 tile is used as test >> board. Verifications are running with libhugetlbfs and ltp. > I haven't reviewed your patches yet but just a few thoughts. Do you > have a clear case where 16MB super-sections is needed? Do you have > memory beyond 4G that can only be accessed with super-sections? This > is an optional feature and not all processors have it. The initial support only focus on ARMv7 based processor like cortex A9, basically that's what I got in my hand, one Versatile Express Cortex-A9x4 tile board. > It also doesn't > help much with the TLB hit rate since the micro-TLB most likely only > support 1MB sections. > I am afraid I can't agree with you on this. Truly there is no specific statement about whether micro-TLB support 16MB super-sections, but main-TLB actually helps the hit rate, and main-TLB support all page size mapping. > BTW, I have a (simpler) implementation of hugetlbfs with 2MB sections > but for LPAE only: > > http://git.kernel.org/?p=linux/kernel/git/cmarinas/linux-arm-arch.git;a=shortlog;h=refs/heads/hugetlb > Thanks for your information, I will study your code carefully :) -- I am a slow learner but I will keep trying to fight for my dreams! --bill