public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Tan Xiaojun <tanxiaojun@huawei.com>
To: <catalin.marinas@arm.com>, <will.deacon@arm.com>,
	<hanjun.guo@linaro.org>, <xieyisheng1@huawei.com>,
	<tglx@linutronix.de>, <punit.agrawal@arm.com>,
	<boris.ostrovsky@oracle.com>
Cc: <linux-kernel@vger.kernel.org>,
	"linux-arm-kernel@lists.infradead.org" 
	<linux-arm-kernel@lists.infradead.org>
Subject: Re: [bug report & help] arm64: ltp testcase "migrate_pages01" failed
Date: Tue, 17 Oct 2017 10:58:53 +0800	[thread overview]
Message-ID: <59E571ED.9090106@huawei.com> (raw)
In-Reply-To: <59E49B2B.8020802@huawei.com>

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.
> 
> 
> .
> 

  reply	other threads:[~2017-10-17  3:03 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-10-16 11:42 [bug report & help] arm64: ltp testcase "migrate_pages01" failed Tan Xiaojun
2017-10-17  2:58 ` Tan Xiaojun [this message]
2017-10-17  9:23   ` Will Deacon
2017-10-17 11:33     ` Tan Xiaojun
2017-10-17 13:19     ` Yisheng Xie
2017-10-18  1:45       ` Yisheng Xie

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=59E571ED.9090106@huawei.com \
    --to=tanxiaojun@huawei.com \
    --cc=boris.ostrovsky@oracle.com \
    --cc=catalin.marinas@arm.com \
    --cc=hanjun.guo@linaro.org \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=punit.agrawal@arm.com \
    --cc=tglx@linutronix.de \
    --cc=will.deacon@arm.com \
    --cc=xieyisheng1@huawei.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox