Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754207AbcLLJyk (ORCPT ); Mon, 12 Dec 2016 04:54:40 -0500 Received: from szxga02-in.huawei.com ([119.145.14.65]:19992 "EHLO szxga02-in.huawei.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751574AbcLLJyi (ORCPT ); Mon, 12 Dec 2016 04:54:38 -0500 Subject: Re: [PATCH] arm64: mm: Fix NOMAP page initialization To: Robert Richter , Russell King , Catalin Marinas , "Will Deacon" References: <1481307042-29773-1-git-send-email-rrichter@cavium.com> <83d6e6d0-cfb3-ec8b-241b-ec6a50dc2aa9@huawei.com> CC: Ard Biesheuvel , David Daney , Mark Rutland , Hanjun Guo , James Morse , , , , Hanjun Guo , Xishi Qiu From: Yisheng Xie Message-ID: <9168b603-04aa-4302-3197-00f17fb336bd@huawei.com> Date: Mon, 12 Dec 2016 17:53:02 +0800 User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:45.0) Gecko/20100101 Thunderbird/45.1.0 MIME-Version: 1.0 In-Reply-To: <83d6e6d0-cfb3-ec8b-241b-ec6a50dc2aa9@huawei.com> Content-Type: text/plain; charset="windows-1252" Content-Transfer-Encoding: 7bit X-Originating-IP: [10.177.29.40] X-CFilter-Loop: Reflected Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 2190 Lines: 64 hi Robert, On 2016/12/12 11:12, Yisheng Xie wrote: > hi Robert, > > On 2016/12/10 2:10, Robert Richter wrote: >> On ThunderX systems with certain memory configurations we see the >> following BUG_ON(): >> >> kernel BUG at mm/page_alloc.c:1848! >> >> This happens for some configs with 64k page size enabled. The BUG_ON() >> checks if start and end page of a memmap range belongs to the same >> zone. >> >> The BUG_ON() check fails if a memory zone contains NOMAP regions. In >> this case the node information of those pages is not initialized. This >> causes an inconsistency of the page links with wrong zone and node >> information for that pages. NOMAP pages from node 1 still point to the >> mem zone from node 0 and have the wrong nid assigned. >> > The patch can work for zone contains NOMAP regions. > > However, if BIOS do not add WB/WT/WC attribute to a physical address range, the > is_memory(md) will return false and this range will not be added to memblock. > efi_init > -> reserve_regions > if (is_memory(md)) { > early_init_dt_add_memory_arch(paddr, size); > > if (!is_usable_memory(md)) > memblock_mark_nomap(paddr, size); > } > > Then BUG_ON() check will also fails. Any idea about it? > It seems that memblock_is_memory() is also too strict for early_pfn_valid, so what about this patch, which use common pfn_valid as early_pfn_valid when CONFIG_HAVE_ARCH_PFN_VALID=y: ------------ diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h index 0f088f3..9d596f3 100644 --- a/include/linux/mmzone.h +++ b/include/linux/mmzone.h @@ -1200,7 +1200,17 @@ static inline int pfn_present(unsigned long pfn) #define pfn_to_nid(pfn) (0) #endif +#ifdef CONFIG_HAVE_ARCH_PFN_VALID +static inline int early_pfn_valid(unsigned long pfn) +{ + if (pfn_to_section_nr(pfn) >= NR_MEM_SECTIONS) + return 0; + return valid_section(__nr_to_section(pfn_to_section_nr(pfn))); +} +#define early_pfn_valid early_pfn_valid +#else #define early_pfn_valid(pfn) pfn_valid(pfn) +#endif void sparse_init(void); #else #define sparse_init() do {} while (0)