_fiemap_filter awk brokeness causing test failures

_fiemap_filter awk brokeness causing test failures

[Nikolay Borisov]

Hello,

Running the script at http://paste.opensuse.org/6056e809 produces the
following output:

root@ubuntu-virtual:~/xfstests-dev# ./test.sh /media/test/file1
0: [0..39]: hole

However, if I run the xfs_io -c "fiemap -v" command I get something
different:

root@ubuntu-virtual:~/xfstests-dev# xfs_io -c "fiemap -v" /media/test/file1
/media/test/file1:
 EXT: FILE-OFFSET      BLOCK-RANGE      TOTAL FLAGS
   0: [0..7]:          hole                 8
   1: [8..23]:         192..207            16 0x801
   2: [24..39]:        hole                16


Finally, if I remove the invocation of _coalesce_extents in the script I
get:
root@ubuntu-virtual:~/xfstests-dev# ./test.sh /media/test/file1
0: [0..7]: hole
2: [24..39]: hole


The script is a streamlined version of _test_generic_punc into a whole.
However, due to the wrong output it's causing this particular test to
fail and some other which deal with hole punching. The correct output
should be :

0: [0..7]: hole
1: [8..23]: unwritten
2: [24..39]: hole

My AWK version is the default coming with ubuntu 16.04:

awk -W version
mawk 1.3.3 Nov 1996, Copyright (C) Michael D. Brennan

compiled limits:
max NF             32767
sprintf buffer      2040

However, if I install gawk the filter works as expected. Perhaps gawk
should be made a hard requirement of xfstests?

Re: _fiemap_filter awk brokeness causing test failures

[Eryu Guan]

Hi,

On Wed, Mar 01, 2017 at 06:48:04PM +0200, Nikolay Borisov wrote:
> Hello,
> 
> Running the script at http://paste.opensuse.org/6056e809 produces the
> following output:
> 
> root@ubuntu-virtual:~/xfstests-dev# ./test.sh /media/test/file1
> 0: [0..39]: hole
> 
> However, if I run the xfs_io -c "fiemap -v" command I get something
> different:
> 
> root@ubuntu-virtual:~/xfstests-dev# xfs_io -c "fiemap -v" /media/test/file1
> /media/test/file1:
>  EXT: FILE-OFFSET      BLOCK-RANGE      TOTAL FLAGS
>    0: [0..7]:          hole                 8
>    1: [8..23]:         192..207            16 0x801
>    2: [24..39]:        hole                16
> 
> 
> Finally, if I remove the invocation of _coalesce_extents in the script I
> get:
> root@ubuntu-virtual:~/xfstests-dev# ./test.sh /media/test/file1
> 0: [0..7]: hole
> 2: [24..39]: hole
> 
> 
> The script is a streamlined version of _test_generic_punc into a whole.
> However, due to the wrong output it's causing this particular test to
> fail and some other which deal with hole punching. The correct output
> should be :
> 
> 0: [0..7]: hole
> 1: [8..23]: unwritten
> 2: [24..39]: hole
> 
> My AWK version is the default coming with ubuntu 16.04:

I haven't looked into these two filters deeply, but a quick test showed
that mawk works fine for me too. But the mawk version is 1.3.4, and it's
from RHEL7 epel repo.

# mawk -W version
mawk 1.3.4 20131226
Copyright 2013, Thomas E. Dickey
Copyright 1996, Michael D. Brennan

internal regex
compiled limits:
max NF             32767
sprintf buffer      2040

Perhaps that's an mawk bug caused the test failure?

Thanks,
Eryu

> 
> awk -W version
> mawk 1.3.3 Nov 1996, Copyright (C) Michael D. Brennan
> 
> compiled limits:
> max NF             32767
> sprintf buffer      2040
> 
> However, if I install gawk the filter works as expected. Perhaps gawk
> should be made a hard requirement of xfstests?
> --
> To unsubscribe from this list: send the line "unsubscribe fstests" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: _fiemap_filter awk brokeness causing test failures

[Nikolay Borisov]

On  2.03.2017 18:32, Eryu Guan wrote:
> Hi,
> 
> On Wed, Mar 01, 2017 at 06:48:04PM +0200, Nikolay Borisov wrote:
>> Hello,
>>
>> Running the script at http://paste.opensuse.org/6056e809 produces the
>> following output:
>>
>> root@ubuntu-virtual:~/xfstests-dev# ./test.sh /media/test/file1
>> 0: [0..39]: hole
>>
>> However, if I run the xfs_io -c "fiemap -v" command I get something
>> different:
>>
>> root@ubuntu-virtual:~/xfstests-dev# xfs_io -c "fiemap -v" /media/test/file1
>> /media/test/file1:
>>  EXT: FILE-OFFSET      BLOCK-RANGE      TOTAL FLAGS
>>    0: [0..7]:          hole                 8
>>    1: [8..23]:         192..207            16 0x801
>>    2: [24..39]:        hole                16
>>
>>
>> Finally, if I remove the invocation of _coalesce_extents in the script I
>> get:
>> root@ubuntu-virtual:~/xfstests-dev# ./test.sh /media/test/file1
>> 0: [0..7]: hole
>> 2: [24..39]: hole
>>
>>
>> The script is a streamlined version of _test_generic_punc into a whole.
>> However, due to the wrong output it's causing this particular test to
>> fail and some other which deal with hole punching. The correct output
>> should be :
>>
>> 0: [0..7]: hole
>> 1: [8..23]: unwritten
>> 2: [24..39]: hole
>>
>> My AWK version is the default coming with ubuntu 16.04:
> 
> I haven't looked into these two filters deeply, but a quick test showed
> that mawk works fine for me too. But the mawk version is 1.3.4, and it's
> from RHEL7 epel repo.
> 
> # mawk -W version
> mawk 1.3.4 20131226
> Copyright 2013, Thomas E. Dickey
> Copyright 1996, Michael D. Brennan
> 
> internal regex
> compiled limits:
> max NF             32767
> sprintf buffer      2040
> 
> Perhaps that's an mawk bug caused the test failure?

Or maybe anything other than plain old awk would work.

> 
> Thanks,
> Eryu
> 
>>
>> awk -W version
>> mawk 1.3.3 Nov 1996, Copyright (C) Michael D. Brennan
>>
>> compiled limits:
>> max NF             32767
>> sprintf buffer      2040
>>
>> However, if I install gawk the filter works as expected. Perhaps gawk
>> should be made a hard requirement of xfstests?
>> --
>> To unsubscribe from this list: send the line "unsubscribe fstests" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>