All of lore.kernel.org
 help / color / mirror / Atom feed
From: Cyril Hrubis <chrubis@suse.cz>
To: ltp@lists.linux.it
Subject: [LTP] [PATCH 3/4] memcg_lib.sh: Get rid of sleep 1 in signal_memcg_process
Date: Mon, 5 Sep 2016 14:50:20 +0200	[thread overview]
Message-ID: <20160905125020.GC14029@rei.lan> (raw)
In-Reply-To: <57C94968.5070609@oracle.com>

Hi!
> > This further brings down the test runtime down by another minute, now
> > the tests finish in about 30 second while we also assert that the
> > usage_in_bytes gets incremented as process consumes memory.
> 
> 50s vs 2m30s in my environment. Cool :)

:)

> >  signal_memcg_process()
> >  {
> > -	pid=$1
> > +	local pid=$1
> > +	local size=$2
> > +	local path=$3
> > +	local usage_start=$(cat ${path}memory.usage_in_bytes)
> 
> A side note.
> 
> In fact, in LTP we use 'local' in test cases written for '/bin/sh',
> whereas 'local' is not in POSIX. Maybe document this fact in the test
> writing guidelines?

Sure. FYI Debian has similar policy see 10.4 in
https://www.debian.org/doc/debian-policy/ch-files.html

> > +
> >  	kill -s USR1 $pid 2> /dev/null
> > -	sleep 1
> > +
> > +	if [ -z "$size" ]; then
> > +		return
> > +	fi
> > +
> > +	while kill -0 $pid 2> /dev/null; do
> > +		local usage=$(cat ${path}memory.usage_in_bytes)
> > +		local diff_a=$((usage_start - usage))
> > +		local diff_b=$((usage - usage_start))
> > +
> > +		if [ "$diff_a" -ge "$size" -o "$diff_b" -ge "$size" ]; then
> > +			return
> > +		fi
> > +
> > +		tst_sleep 100ms
> > +	done
> 
> Maybe add some limit on the number of checks? In case the process
> consumes less memory than in $size.

Will do.

> >  # Test hierarchical_memory_limit with enabling hierarchical accounting
> > diff --git a/testcases/kernel/controllers/memcg/functional/memcg_usage_in_bytes_test.sh b/testcases/kernel/controllers/memcg/functional/memcg_usage_in_bytes_test.sh
> > index ab86318..69dc874 100755
> > --- a/testcases/kernel/controllers/memcg/functional/memcg_usage_in_bytes_test.sh
> > +++ b/testcases/kernel/controllers/memcg/functional/memcg_usage_in_bytes_test.sh
> > @@ -34,7 +34,7 @@ TST_TOTAL=2
> >  # Test memory.usage_in_bytes
> >  testcase_1()
> >  {
> > -	test_mem_stat "--mmap-anon" $((PAGESIZE*1024)) \
> > +	test_mem_stat "--mmap-anon" $((PAGESIZE*1024)) $((PAGESIZE*1024)) \
> >  		"memory.usage_in_bytes" $((PAGESIZE*1024)) 0
> >  }
> >  
> > @@ -48,7 +48,7 @@ testcase_2()
> >  
> >  	echo $((PAGESIZE*2048)) > memory.limit_in_bytes
> >  	echo $((PAGESIZE*2048)) > memory.memsw.limit_in_bytes
> > -	test_mem_stat "--mmap-anon" $((PAGESIZE*1024)) \
> > +	test_mem_stat "--mmap-anon" $((PAGESIZE*1024)) $((PAGESIZE*1024)) \
> >  		"memory.memsw.usage_in_bytes" $((PAGESIZE*1024)) 0
> >  }
> 
> The above two calls need '0' substituted with 'false'.

My bad, I've missed that one.

-- 
Cyril Hrubis
chrubis@suse.cz

  parent reply	other threads:[~2016-09-05 12:50 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-09-01 17:18 [LTP] [PATCH 1/4] memcg_lib.sh: Print test number at test start Cyril Hrubis
2016-09-01 17:18 ` [LTP] [PATCH 2/4] memcg_lib.sh: Move all the kill -s XXX to functions Cyril Hrubis
2016-09-02  9:27   ` Stanislav Kholmanskikh
2016-09-01 17:18 ` [LTP] [PATCH 3/4] memcg_lib.sh: Get rid of sleep 1 in signal_memcg_process Cyril Hrubis
2016-09-02  9:42   ` Stanislav Kholmanskikh
2016-09-05 10:18     ` Jan Stancek
2016-09-05 12:10       ` Cyril Hrubis
2016-09-05 12:50     ` Cyril Hrubis [this message]
2016-09-07 10:19     ` Cyril Hrubis
2016-09-08  8:38       ` Stanislav Kholmanskikh
2016-09-01 17:18 ` [LTP] [PATCH 4/4] memcg_function_test.sh: Split Cyril Hrubis
2016-09-02 10:28   ` Stanislav Kholmanskikh

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=20160905125020.GC14029@rei.lan \
    --to=chrubis@suse.cz \
    --cc=ltp@lists.linux.it \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.