From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1757186AbdJQDDp (ORCPT ); Mon, 16 Oct 2017 23:03:45 -0400 Received: from szxga04-in.huawei.com ([45.249.212.190]:8461 "EHLO szxga04-in.huawei.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751757AbdJQDDo (ORCPT ); Mon, 16 Oct 2017 23:03:44 -0400 Subject: Re: [bug report & help] arm64: ltp testcase "migrate_pages01" failed CC: , "linux-arm-kernel@lists.infradead.org" References: <59E49B2B.8020802@huawei.com> From: Tan Xiaojun To: , , , , , , Message-ID: <59E571ED.9090106@huawei.com> Date: Tue, 17 Oct 2017 10:58:53 +0800 User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:38.0) Gecko/20100101 Thunderbird/38.6.0 MIME-Version: 1.0 In-Reply-To: <59E49B2B.8020802@huawei.com> Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 7bit X-Originating-IP: [10.177.21.79] X-CFilter-Loop: Reflected X-Mirapoint-Virus-RAPID-Raw: score=unknown(0), refid=str=0001.0A020202.59E57248.00EA,ss=1,re=0.000,recu=0.000,reip=0.000,cl=1,cld=1,fgs=0, ip=0.0.0.0, so=2014-11-16 11:51:01, dmn=2013-03-21 17:37:32 X-Mirapoint-Loop-Id: 4052984047ee916d159a1a9d58064849 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Hi all, I'm not sure if this is the problem on arm64 numa. What do you think ? By the way, this testcase can be successful in any case on x86. Thanks. Xiaojun. On 2017/10/16 19:42, Tan Xiaojun wrote: > Hi all, > > I test ltp in Hisilicon D05 board and get a failed result about the testcase "migrate_pages01". > > In fact, The sub testcase "test_invalid_nodes" failed. The testcase is to find a invalid numa node and migrate memory pages to this node via syscall of "migrate_pages". > The expected result of this case is returning "-1", but it actually return "0". > > -------------------------------------------------------- > # ./migrate_pages01 > migrate_pages01 0 TINFO : test_empty_mask > migrate_pages01 1 TPASS : expected ret success: returned value = 0 > migrate_pages01 0 TINFO : test_invalid_pid -1 > migrate_pages01 2 TPASS : expected ret success: returned value = -1 > migrate_pages01 3 TPASS : expected failure: TEST_ERRNO=ESRCH(3): No such process > migrate_pages01 0 TINFO : test_invalid_pid unused pid > migrate_pages01 4 TPASS : expected ret success: returned value = -1 > migrate_pages01 5 TPASS : expected failure: TEST_ERRNO=ESRCH(3): No such process > migrate_pages01 0 TINFO : test_invalid_masksize > migrate_pages01 6 TPASS : expected ret success: returned value = -1 > migrate_pages01 7 TPASS : expected failure: TEST_ERRNO=EINVAL(22): Invalid argument > migrate_pages01 0 TINFO : test_invalid_mem -1 > migrate_pages01 8 TPASS : expected ret success: returned value = -1 > migrate_pages01 9 TPASS : expected failure: TEST_ERRNO=EFAULT(14): Bad address > migrate_pages01 0 TINFO : test_invalid_mem invalid prot > migrate_pages01 10 TPASS : expected ret success: returned value = -1 > migrate_pages01 11 TPASS : expected failure: TEST_ERRNO=EFAULT(14): Bad address > migrate_pages01 0 TINFO : test_invalid_mem unmmaped > migrate_pages01 12 TPASS : expected ret success: returned value = -1 > migrate_pages01 13 TPASS : expected failure: TEST_ERRNO=EFAULT(14): Bad address > migrate_pages01 0 TINFO : test_invalid_nodes > migrate_pages01 14 TFAIL : migrate_pages_common.c:45: unexpected failure - returned value = 0, expected: -1 > migrate_pages01 15 TFAIL : migrate_pages_common.c:55: call succeeded unexpectedly > migrate_pages01 0 TINFO : test_invalid_perm > migrate_pages01 16 TPASS : expected ret success: returned value = -1 > migrate_pages01 17 TPASS : expected failure: TEST_ERRNO=EPERM(1): Operation not permitted > -------------------------------------------------------- > > I debug and find a interesting thing, this case does not always fail. > > 1) If one or several numa nodes have no memory, this case will run successfully like below: > > -------------------- > available: 4 nodes (0-3) > node 0 cpus: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 > node 0 size: 65309 MB > node 0 free: 61650 MB > node 1 cpus: 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 > node 1 size: 65404 MB > node 1 free: 61377 MB > node 2 cpus: 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 > node 2 size: 65401 MB > node 2 free: 62316 MB > node 3 cpus: 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 > node 3 size: 0 MB > node 3 free: 0 MB > node distances: > node 0 1 2 3 > 0: 10 15 20 20 > 1: 15 10 20 20 > 2: 20 20 10 15 > 3: 20 20 15 10 > --------------------- > > This testcase will find node number 3 and migrate pages to node 3. And syscall of "migrate_pages" return -1, test succeeded. > > 2) In most cases, all nodes have memory, and the testcase will get non-existent node like node number 4. The syscall of "migrate_pages" should also return -1, but return 0 actually. > So the testcase failed. > > I think it is a problem in arm64. But I am not familiar with numa, so I ask for help from you. > > Thanks. > Xiaojun. > > > . >