review: xfstests/192 atime test

review: xfstests/192 atime test

[Timothy Shimmin]

[-- Attachment #1: Type: text/plain, Size: 95 bytes --]

Attached simple atime test.
192.out and group file would be added appropriately.
Thanks.
--Tim

[-- Attachment #2: 192 --]
[-- Type: text/plain, Size: 1285 bytes --]

#! /bin/sh
# FS QA Test No. 192
#
# Simple test of atime - ensure it is persistent after unmount
#
#-----------------------------------------------------------------------
# Copyright (c) 2008 Silicon Graphics, Inc.  All Rights Reserved.
#-----------------------------------------------------------------------
#
# creator
owner=tes@emu.melbourne.sgi.com

seq=`basename $0`
echo "QA output created by $seq"

here=`pwd`
tmp=/tmp/$$
status=1	# failure is the default!
trap "exit \$status" 0 1 2 3 15

_access_time()
{
	stat --format=%X $1
}

# get standard environment, filters and checks
. ./common.rc
. ./common.filter

# real QA test starts here

_supported_fs xfs udf nfs
_supported_os Linux
delay=150

testfile=$TEST_DIR/testfile
rm -f $testfile
rm -f $seq.full

echo test >$testfile
time1=`_access_time $testfile | tee -a $seq.full` 

echo "sleep for $delay"
sleep $delay # sleep to allow time to move on for access
cat $testfile
time2=`_access_time $testfile | tee -a $seq.full` 

cd /
umount $TEST_DIR
mount $TEST_DIR
time3=`_access_time $testfile | tee -a $seq.full`

delta1=`expr $time2 - $time1`
delta2=`expr $time3 - $time1`

echo "delta1 - access time after sleep in-core: $delta1"
echo "delta2 - access time after sleep on-disk: $delta2"

# success, all done
status=0
exit

Re: review: xfstests/192 atime test

[Dave Chinner]

On Tue, Nov 11, 2008 at 05:11:16PM +1100, Timothy Shimmin wrote:
> Attached simple atime test.
> 192.out and group file would be added appropriately.
> Thanks.
> --Tim

> _supported_fs xfs udf nfs
> _supported_os Linux
> delay=150

150s sleep? It only needs a few seconds for this test.

> time3=`_access_time $testfile | tee -a $seq.full`
> 
> delta1=`expr $time2 - $time1`
> delta2=`expr $time3 - $time1`
> 
> echo "delta1 - access time after sleep in-core: $delta1"
> echo "delta2 - access time after sleep on-disk: $delta2"

Why bother with deltas? if the numbers are the same, just output
"same", if they are different, output "different"....

FWIW, is there a patch being proposed that fixes this problem?

Cheers,

Dave.
-- 
Dave Chinner
david@fromorbit.com

Re: review: xfstests/192 atime test

[Timothy Shimmin]

Dave Chinner wrote:
> On Tue, Nov 11, 2008 at 05:11:16PM +1100, Timothy Shimmin wrote:
>> Attached simple atime test.
>> 192.out and group file would be added appropriately.
>> Thanks.
>> --Tim
> 
>> _supported_fs xfs udf nfs
>> _supported_os Linux
>> delay=150
> 
> 150s sleep? It only needs a few seconds for this test.
> 
Not for me. Passed every time at a few secs,
that is why I tried longer as was originally reported.

>> time3=`_access_time $testfile | tee -a $seq.full`
>>
>> delta1=`expr $time2 - $time1`
>> delta2=`expr $time3 - $time1`
>>
>> echo "delta1 - access time after sleep in-core: $delta1"
>> echo "delta2 - access time after sleep on-disk: $delta2"
> 
> Why bother with deltas? if the numbers are the same, just output
> "same", if they are different, output "different"....
> 
Fine, will do.

> FWIW, is there a patch being proposed that fixes this problem?
> 
I lost the thread with people to'ing and fro'ing ;-)
I'd have to relook at the last emails.

--Tim

Re: review: xfstests/192 atime test

[Timothy Shimmin]

Timothy Shimmin wrote:
> Dave Chinner wrote:
>> On Tue, Nov 11, 2008 at 05:11:16PM +1100, Timothy Shimmin wrote:
>>> Attached simple atime test.
>>> 192.out and group file would be added appropriately.
>>> Thanks.
>>> --Tim
>>> _supported_fs xfs udf nfs
>>> _supported_os Linux
>>> delay=150
>> 150s sleep? It only needs a few seconds for this test.
>>
> Not for me. Passed every time at a few secs,
> that is why I tried longer as was originally reported.
> 
>>> time3=`_access_time $testfile | tee -a $seq.full`
>>>
>>> delta1=`expr $time2 - $time1`
>>> delta2=`expr $time3 - $time1`
>>>
>>> echo "delta1 - access time after sleep in-core: $delta1"
>>> echo "delta2 - access time after sleep on-disk: $delta2"
>> Why bother with deltas? if the numbers are the same, just output
>> "same", if they are different, output "different"....
>>
> Fine, will do.
> 
The only other thing is that I am not just testing the bug
as it stands - i.e. if the atime is being lost on-disk.
I want to test that the atime is updated appropriately.

--Tim

Re: review: xfstests/192 atime test

[Timothy Shimmin]

Timothy Shimmin wrote:
> Timothy Shimmin wrote:
>> Dave Chinner wrote:
>>> On Tue, Nov 11, 2008 at 05:11:16PM +1100, Timothy Shimmin wrote:
>>>> Attached simple atime test.
>>>> 192.out and group file would be added appropriately.
>>>> Thanks.
>>>> --Tim
>>>> _supported_fs xfs udf nfs
>>>> _supported_os Linux
>>>> delay=150
>>> 150s sleep? It only needs a few seconds for this test.
>>>
>> Not for me. Passed every time at a few secs,
>> that is why I tried longer as was originally reported.
>>
>>>> time3=`_access_time $testfile | tee -a $seq.full`
>>>>
>>>> delta1=`expr $time2 - $time1`
>>>> delta2=`expr $time3 - $time1`
>>>>
>>>> echo "delta1 - access time after sleep in-core: $delta1"
>>>> echo "delta2 - access time after sleep on-disk: $delta2"
>>> Why bother with deltas? if the numbers are the same, just output
>>> "same", if they are different, output "different"....
>>>
>> Fine, will do.
>>
> The only other thing is that I am not just testing the bug
> as it stands - i.e. if the atime is being lost on-disk.
> I want to test that the atime is updated appropriately.
> 
> --Tim
> 

Ping?
Not fussed?

Thanks,
--Tim

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs