public inbox for ltp@lists.linux.it
 help / color / mirror / Atom feed
From: Zhouping Liu <zliu@redhat.com>
To: Wanlong Gao <gaowanlong@cn.fujitsu.com>,
	Jan Stancek <jstancek@redhat.com>,
	Caspar Zhang <czhang@redhat.com>
Cc: LTP List <ltp-list@lists.sourceforge.net>
Subject: Re: [LTP] [PATCH] mem/oom: fixed a cpuset error
Date: Fri, 26 Apr 2013 22:20:04 -0400 (EDT)	[thread overview]
Message-ID: <879965201.3279983.1367029204685.JavaMail.root@redhat.com> (raw)
In-Reply-To: <78191d90d19975e03308f3c05d0bf66fd1471ba5.1366772430.git.zliu@redhat.com>



----- Original Message -----
> From: "Zhouping Liu" <zliu@redhat.com>
> To: "LTP List" <ltp-list@lists.sourceforge.net>
> Sent: Wednesday, April 24, 2013 11:02:04 AM
> Subject: [LTP] [PATCH] mem/oom: fixed a cpuset error
> 
> For the below special NUMA system, oom0[4|5] failed:
>  # numactl -H
>  available: 2 nodes (0-1)
>  node 0 cpus: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
>  24 25 26 27
>  node 0 size: 0 MB
>  node 0 free: 0 MB
>  node 1 cpus:
>  node 1 size: 16384 MB
>  node 1 free: 14173 MB
>  node distances:
>  node   0   1
>    0:  10  40
>    1:  40  10
> 
> failed log:
>  oom04       1  TBROK  :  write /dev/cpuset/1/cpuset.mems: errno=EINVAL(22):
>  Invalid argument
> 
> The reason is that node0 only contains all CPUs, no any memory,
> and node1 contains all memory, but no CPUs, in the previous codes,
> we only do cpuset testing on a independent node, which caused the
> sub-cpuset only contains CPUs or memory in one node, that's not
> permitted. The patch fixed it.

Wanlong, Jan and Caspar,

any comments for this patch?

Thanks,
Zhouping

> 
> Signed-off-by: Zhouping Liu <zliu@redhat.com>
> 
> Tested on serveral machines, all PASSed.
> ---
>  testcases/kernel/mem/lib/mem.c   | 13 ++++++++++++-
>  testcases/kernel/mem/oom/oom04.c | 19 +++++++++++++------
>  testcases/kernel/mem/oom/oom05.c | 25 +++++++++++++++----------
>  3 files changed, 40 insertions(+), 17 deletions(-)
> 
> diff --git a/testcases/kernel/mem/lib/mem.c b/testcases/kernel/mem/lib/mem.c
> index 5397177..b2ad562 100644
> --- a/testcases/kernel/mem/lib/mem.c
> +++ b/testcases/kernel/mem/lib/mem.c
> @@ -792,7 +792,18 @@ void write_cpusets(long nd)
>  	write_cpuset_files(CPATH_NEW, "mems", buf);
>  
>  	gather_node_cpus(cpus, nd);
> -	write_cpuset_files(CPATH_NEW, "cpus", cpus);
> +	/*
> +	 * If the 'nd' node didn't contains any CPUs,
> +	 * the CPU0 will be used as the cpuset.cpus.
> +	 */
> +	if (strlen(cpus) != 0) {
> +		write_cpuset_files(CPATH_NEW, "cpus", cpus);
> +	} else {
> +		tst_resm(TINFO, "No any CPUs in node%ld", nd);
> +		tst_resm(TINFO, "Using CPU0 to cpuset.cpus "
> +				"for the special scenario");
> +		write_cpuset_files(CPATH_NEW, "cpus", "0");
> +	}
>  
>  	SAFE_FILE_PRINTF(NULL, CPATH_NEW "/tasks", "%d", getpid());
>  }
> diff --git a/testcases/kernel/mem/oom/oom04.c
> b/testcases/kernel/mem/oom/oom04.c
> index 4d3f2f4..40c1198 100644
> --- a/testcases/kernel/mem/oom/oom04.c
> +++ b/testcases/kernel/mem/oom/oom04.c
> @@ -85,6 +85,8 @@ int main(int argc, char *argv[])
>  
>  void setup(void)
>  {
> +	int memnode, ret;
> +
>  	tst_require_root(NULL);
>  	tst_sig(FORK, DEF_HANDLER, cleanup);
>  	TEST_PAUSE;
> @@ -93,12 +95,17 @@ void setup(void)
>  	set_sys_tune("overcommit_memory", 1, 1);
>  
>  	mount_mem("cpuset", "cpuset", NULL, CPATH, CPATH_NEW);
> -	if (is_numa(cleanup) > 0)
> -		/* For NUMA system, using the first node for cpuset.mems */
> -		write_cpusets(get_a_numa_node(cleanup));
> -	else
> -		/* For nonNUMA system, using node0 for cpuset.mems */
> -		write_cpusets(0);
> +
> +	/*
> +	 * Not any nodes contains memory, so using get_allowed_nodes()
> +	 * to get a memory node, the operation is also appropriate for
> +	 * NonNUMA system.
> +	 */
> +	ret = get_allowed_nodes(NH_MEMS, 1, &memnode);
> +	if (ret < 0)
> +		tst_brkm(TBROK, NULL, "Got one memory node failed "
> +				      "using get_allowed_nodes()");
> +	write_cpusets(memnode);
>  }
>  
>  void cleanup(void)
> diff --git a/testcases/kernel/mem/oom/oom05.c
> b/testcases/kernel/mem/oom/oom05.c
> index 15feba5..520d657 100644
> --- a/testcases/kernel/mem/oom/oom05.c
> +++ b/testcases/kernel/mem/oom/oom05.c
> @@ -108,24 +108,29 @@ int main(int argc, char *argv[])
>  
>  void setup(void)
>  {
> +	int ret, memnode;
> +
>  	tst_require_root(NULL);
>  	tst_sig(FORK, DEF_HANDLER, cleanup);
>  	TEST_PAUSE;
>  
> +	overcommit = get_sys_tune("overcommit_memory");
> +	set_sys_tune("overcommit_memory", 1, 1);
> +
>  	mount_mem("memcg", "cgroup", "memory", MEMCG_PATH, MEMCG_PATH_NEW);
>  	mount_mem("cpuset", "cpuset", NULL, CPATH, CPATH_NEW);
>  	write_memcg();
>  
> -	set_sys_tune("overcommit_memory", 1, 1);
> -
> -	if (is_numa(cleanup))
> -		/* For NUMA system, using the first node for cpuset.mems */
> -		write_cpusets(get_a_numa_node(cleanup));
> -	else
> -		/* For nonNUMA system, using node0 for cpuset.mems */
> -		write_cpusets(0);
> -
> -	overcommit = get_sys_tune("overcommit_memory");
> +        /*
> +	 * Not any nodes contains memory, so using get_allowed_nodes()
> +	 * to get a memory node, the operation is also appropriate for
> +	 * NonNUMA system.
> +	 */
> +	ret = get_allowed_nodes(NH_MEMS, 1, &memnode);
> +	if (ret < 0)
> +		tst_brkm(TBROK, NULL, "Got one memory node failed "
> +				      "using get_allowed_nodes()");
> +	write_cpusets(memnode);
>  }
>  
>  void cleanup(void)
> --
> 1.7.11.7
> 
> 
> ------------------------------------------------------------------------------
> Try New Relic Now & We'll Send You this Cool Shirt
> New Relic is the only SaaS-based application performance monitoring service
> that delivers powerful full stack analytics. Optimize and monitor your
> browser, app, & servers with just a few lines of code. Try New Relic
> and get this awesome Nerd Life shirt! http://p.sf.net/sfu/newrelic_d2d_apr
> _______________________________________________
> Ltp-list mailing list
> Ltp-list@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/ltp-list
> 

------------------------------------------------------------------------------
Try New Relic Now & We'll Send You this Cool Shirt
New Relic is the only SaaS-based application performance monitoring service 
that delivers powerful full stack analytics. Optimize and monitor your
browser, app, & servers with just a few lines of code. Try New Relic
and get this awesome Nerd Life shirt! http://p.sf.net/sfu/newrelic_d2d_apr
_______________________________________________
Ltp-list mailing list
Ltp-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ltp-list

      reply	other threads:[~2013-04-27  2:20 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-04-24  3:02 [LTP] [PATCH] mem/oom: fixed a cpuset error Zhouping Liu
2013-04-27  2:20 ` Zhouping Liu [this message]

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=879965201.3279983.1367029204685.JavaMail.root@redhat.com \
    --to=zliu@redhat.com \
    --cc=czhang@redhat.com \
    --cc=gaowanlong@cn.fujitsu.com \
    --cc=jstancek@redhat.com \
    --cc=ltp-list@lists.sourceforge.net \
    /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