GNU bug report logs - #6131
[PATCH]: fiemap support for efficient sparse file copy

Previous Next

Full log

Message #50 received at submit <at> debbugs.gnu.org (full text, mbox):

From: "jeff.liu" <jeff.liu <at> oracle.com>
To: Jim Meyering <jim <at> meyering.net>
Cc: Sunil Mushran <sunil.mushran <at> oracle.com>, Tao Ma <tao.ma <at> oracle.com>,
	bug-coreutils <at> gnu.org, Joel Becker <Joel.Becker <at> oracle.com>,
	Chris Mason <chris.mason <at> oracle.com>
Subject: Re: bug#6131: [PATCH]: fiemap support for efficient sparse file copy
Date: Mon, 24 May 2010 17:34:01 +0800

jeff.liu wrote:
> Jim Meyering wrote:
>> jeff.liu wrote:
>>> Jim Meyering wrote:
>>>> jeff.liu wrote:
>>>> ...
>>>>>>> Subject: [PATCH 1/1] tests: add a new test for FIEMAP-copy
>>>>>>>
>>>>>>> * tests/cp/sparse-fiemap: Add a new test for FIEMAP-copy against a
>>>>>>> loopbacked ext4 partition.
>>>>>>> * tests/Makefile.am (sparse-fiemap): Reference the new test.
>>>> BTW, I've just made this additional change to your test,
>>>>
>>>> diff --git a/tests/cp/sparse-fiemap b/tests/cp/sparse-fiemap
>>>> index 6312a4c..bdc7ded 100755
>>>> --- a/tests/cp/sparse-fiemap
>>>> +++ b/tests/cp/sparse-fiemap
>>>> @@ -27,6 +27,7 @@ require_root_
>>>>  cwd=`pwd`
>>>>  cleanup_() { cd /; umount "$cwd/mnt"; }
>>>>
>>>> +skip=0
>> More fixes for the test:
>> - remove useless rm
>> - $sparse and $fiemap were not defined in that final test,
>>     so cd into mnt/ and just use the file names instead.
>> - I prefer $(...) to `...`, and it's portable in this context
>>
>> diff --git a/tests/cp/sparse-fiemap b/tests/cp/sparse-fiemap
>> index 32ca5fc..cec5224 100755
>> --- a/tests/cp/sparse-fiemap
>> +++ b/tests/cp/sparse-fiemap
>> @@ -40,18 +40,18 @@ test -s mnt/f                                  || skip=1
>>  test $skip = 1 &&
>>    skip_test_ "insufficient mount/ext4 support"
>>
>> -rm -f mnt/f
>> -
>>  # Create a 1TiB sparse file
>>  dd if=/dev/zero of=mnt/sparse bs=1k count=1 seek=1G || framework_failure
>>
>> +cd mnt || fail=1
>> +
>>  # It takes many minutes to copy this sparse file using the old method.
>>  # By contrast, it takes far less than 1 second using FIEMAP-copy.
>> -timeout 10 cp --sparse=always mnt/sparse mnt/sparse_fiemap || fail=1
>> +timeout 10 cp --sparse=always sparse fiemap || fail=1
>>
>>  # Ensure that the sparse file copied through fiemap has the same size
>>  # in bytes as the original.
>> -test `stat --printf %s $sparse` = `stat --printf %s $fiemap` || fail=1
>> +test $(stat --printf %s sparse) = $(stat --printf %s fiemap) || fail=1
>>
>>  # =================================================
>>  # Ensure that we exercise the FIEMAP-copying code enough
>> --
>> 1.7.1.262.g5ef3d
> 
> Thanks for the info.
> 
> For the point of *- I prefer $(...) to `...`, and it's portable in this context*
> 
> Could you check the tiny patch below, it does the same thing to improve the portability of
> 'tests/cp/sparse', I have referred to it as a model.
> 
> From 9fbffcd0f4d4706f4b88e7ac73ca13b62c789047 Mon Sep 17 00:00:00 2001
> From: Jie Liu <jeff.liu <at> oracle.com>
> Date: Mon, 24 May 2010 16:01:06 +0800
> Subject: [PATCH 1/1] cp: improve the portability of test
> 
> * tests/cp/sparse: improve the portability using shell constructs.
> 
> Signed-off-by: Jie Liu <jeff.liu <at> oracle.com>
> ---
>  tests/cp/sparse |    4 ++--
>  1 files changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/tests/cp/sparse b/tests/cp/sparse
> index 73c2924..dee6de2 100755
> --- a/tests/cp/sparse
> +++ b/tests/cp/sparse
> @@ -28,14 +28,14 @@ require_sparse_support_
>  # It has to be at least 128K in order to be sparse on some systems.
>  # Make its size one larger than 128K, in order to tickle the
>  # bug in coreutils-6.0.
> -size=`expr 128 \* 1024 + 1`
> +size=$((128 * 1024 + 1))
>  dd bs=1 seek=$size of=sparse < /dev/null 2> /dev/null || framework_failure
> 
> 
>  cp --sparse=always sparse copy || fail=1
> 
>  # Ensure that the copy has the same block count as the original.
> -test `stat --printf %b copy` -le `stat --printf %b sparse` || fail=1
> +test $(stat --printf %b copy) -le $(stat --printf %b sparse) || fail=1
> 
>  # Ensure that --sparse={always,never} with --reflink fail.
>  cp --sparse=always --reflink sparse copy && fail=1

Please ignore above patch, I just found another issue in 'tests/cp/sparse', the new created test
file also named to 'sparse', so when it running, the `cp/sparse' will be truncated to `expr 128 \*
1024 + 1`.

Below patch fix it to create a sparse file 'sparse1' instead(I can not find out a better name for
now), s/-le/=/ to compare the block count.

From 0669ac6d0497a3c6abfc5d53202afc6bc47d0d07 Mon Sep 17 00:00:00 2001
From: Jie Liu <jeff.liu <at> oracle.com>
Date: Mon, 24 May 2010 17:29:27 +0800
Subject: [PATCH 1/1] cp: enhance the sparse file copy test

* tests/cp/sparse: fix sparse file name to 'sparse1', improve
the protability using shell constructs.

Signed-off-by: Jie Liu <jeff.liu <at> oracle.com>
---
 tests/cp/sparse |   12 ++++++------
 1 files changed, 6 insertions(+), 6 deletions(-)

diff --git a/tests/cp/sparse b/tests/cp/sparse
index 73c2924..cab8b9c 100755
--- a/tests/cp/sparse
+++ b/tests/cp/sparse
@@ -28,17 +28,17 @@ require_sparse_support_
 # It has to be at least 128K in order to be sparse on some systems.
 # Make its size one larger than 128K, in order to tickle the
 # bug in coreutils-6.0.
-size=`expr 128 \* 1024 + 1`
-dd bs=1 seek=$size of=sparse < /dev/null 2> /dev/null || framework_failure
+size=$((128 * 1024 + 1))
+dd bs=1 seek=$size of=sparse1 < /dev/null 2> /dev/null || framework_failure


-cp --sparse=always sparse copy || fail=1
+cp --sparse=always sparse1 copy || fail=1

 # Ensure that the copy has the same block count as the original.
-test `stat --printf %b copy` -le `stat --printf %b sparse` || fail=1
+test $(stat --printf %b copy) = $(stat --printf %b sparse1) || fail=1

 # Ensure that --sparse={always,never} with --reflink fail.
-cp --sparse=always --reflink sparse copy && fail=1
-cp --sparse=never --reflink sparse copy && fail=1
+cp --sparse=always --reflink sparse1 copy && fail=1
+cp --sparse=never --reflink sparse1 copy && fail=1

 Exit $fail
-- 
1.5.4.3


-- 
With Windows 7, Microsoft is asserting legal control over your computer and is using this power to
abuse computer users.

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.