| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <045d4777-ae52-cf59-bb99-c032eda292d4@suse.com>
Date: Tue, 27 Jun 2017 09:31:47 +0800
From: James Wang <jnwang@...e.com>
To: Omar Sandoval <osandov@...ndov.com>
Cc: osandov@...com, hch@...radead.org, axboe@...com, hare@...e.com,
linux-block@...r.kernel.org, linux-kernel@...r.kernel.org,
mgorman@...e.com
Subject: Re: [PATCH blktests] loop/002: Regression testing for loop device
flush
On 06/27/2017 02:58 AM, Omar Sandoval wrote:
> Hi, James, thanks for sending this in. Sorry for the delay, I've been
> out of the office for a couple of weeks. A few comments below.
>
> On Thu, Jun 08, 2017 at 08:28:12PM +0800, James Wang wrote:
>> Add a regression testing for loop device. when an unbound device
>> be close that take too long time. kernel will consume serveral orders
>> of magnitude more wall time than it does for a mounted device.
>>
>> Signed-off-by: James Wang <jnwang@...e.com>
>> ---
>> tests/loop/002 | 77 ++++++++++++++++++++++++++++++++++++++++++++++++++++++
>> tests/loop/002.out | 2 ++
>> 2 files changed, 79 insertions(+)
>>
>> diff --git a/tests/loop/002 b/tests/loop/002
>> new file mode 100755
>> index 0000000..fd607d1
>> --- /dev/null
>> +++ b/tests/loop/002
>> @@ -0,0 +1,77 @@
>> +#!/bin/bash
>> +#
>> +# Test if close()ing a unbound loop device is too slow
>> +# Copyright (C) 2017 James Wang
>> +#
>> +# This program is free software: you can redistribute it and/or modify
>> +# it under the terms of the GNU General Public License as published by
>> +# the Free Software Foundation, either version 3 of the License, or
>> +# (at your option) any later version.
>> +#
>> +# This program is distributed in the hope that it will be useful,
>> +# but WITHOUT ANY WARRANTY; without even the implied warranty of
>> +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
>> +# GNU General Public License for more details.
>> +#
>> +# You should have received a copy of the GNU General Public License
>> +# along with this program. If not, see <http://www.gnu.org/licenses/>.
>> +
>> +DESCRIPTION="Test if close()ing a unbound loop device is too slow"
>> +
>> +QUICK=1
>> +
>> +function run_test() {
> For consistency with everything else in blktests, please don't use
> "function" when defining a function.
I will fix it.
>> + TIMEFORMAT='%5R'
>> + time {
>> + for f in `ls /dev/loop[0-9]*|sort`; do dd if=$f of=/dev/null bs=512 count=1 >/dev/null 2>&1; done
>> + }
>> +}
>> +function clean_up() {
>> + if lsmod | grep loop >/dev/null 2>&1; then
>> + umount /dev/loop* >/dev/null 2>&1
>> + losetup -D
>> + sleep 5
>> +
>> + if ! rmmod loop;then
>> + return 2;
>> + fi
>> + fi
>> +}
>> +
>> +function prepare() {
>> + modprobe loop max_loop=64
> If loop is already loaded, this won't work, right?
Actually, I could use clean_up() first , but due to My testing machine
has a bug causes clean_up() very slow......
I use call clean_up() before prepare(), make sense?
>
>> + dd if=/dev/zero of=${TMPDIR}/disk bs=512 count=200K >/dev/null 2>&1
>> + for((i=0;i<4;i++))
>> + do
>> + losetup -f ${TMPDIR}/disk;
>> + done
>> + mkfs.ext4 -F /dev/loop0 >/dev/null 2>&1
> Hm, so if I happened to have something I care about on /dev/loop0,
> running blktests will destroy it? This is a no-go.
Yes, but due to our insert loop module and create a fake-disk and bound
to loop0, so format loop0 should doesn't matter.
>> + for((i=0;i<4;i++))
>> + do
>> + mkdir -p t$i;
>> + mount /dev/loop$i t$i;
>> + done
>> +
>> +}
>> +
>> +
>> +test() {
>> + echo "Running ${TEST_NAME}"
>> +
>> + prepare
>> + SECONDS=0
>> + run_test >/dev/null 2>&1
>> + DURATION=${SECONDS}
> Nifty, I didn't know about $SECONDS.
SECONDS is a built-in variable in bash, it will automatic increase.
>
>> +
>> + clean_up
>> + if ! clean_up; then
>> + echo "Test complete"
>> + return 2
>> + fi
>> + echo "Test complete"
>> + if [[ "${DURATION}" -gt 1 ]]; then
>> + return 1
>> + else
>> + return 0
>> + fi
> I'd really like a meaningful output if this test fails, so something
> like this instead of the if/else
>
> if [[ "${DURATION}" -gt 1 ]]; then
> echo "test took too long ($DURATION seconds)"
> fi
I will fix this.
>> +}
>> diff --git a/tests/loop/002.out b/tests/loop/002.out
>> new file mode 100644
>> index 0000000..5c34a37
>> --- /dev/null
>> +++ b/tests/loop/002.out
>> @@ -0,0 +1,2 @@
>> +Running loop/002
>> +Test complete
>> --
>> 2.12.3
>>
> Overall, is there an easier way to test this than setting up 64 loop
> devices at modprobe time? E.g., can you losetup -f and run it on a
> single loop device many times to measure the same issue?
Use many loop devices for get a enough long time to compare with 1 second.
if we only create 1 loop device, I afraid it can't be measured.
In this scenario, I could get the duration of unbound and bound loop
device takes.
OK, I could try your suggestion.
I will send patch later.
James
>
> Thanks again!
>
>
--
SUSE Linux GmbH, GF: Felix Imendörffer, Jane Smithard, Graham Norton, HRB 21284 (AG Nürnberg)
Powered by blists - more mailing lists