fstests.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Yang Xu <xuyang2018.jy@cn.fujitsu.com>
To: "Darrick J. Wong" <darrick.wong@oracle.com>
Cc: Eryu Guan <guaneryu@gmail.com>, <fstests@vger.kernel.org>
Subject: Re: [PATCH v2] common/log: add -l su option in _xfs_log_config
Date: Mon, 24 Feb 2020 10:14:33 +0800	[thread overview]
Message-ID: <484e9bf8-f659-c752-f232-11ee76a353a1@cn.fujitsu.com> (raw)
In-Reply-To: <a801d86e-80ab-b2fb-ba72-721365a8133e@cn.fujitsu.com>



on 2020/02/24 9:51, Yang Xu wrote:
> 
> 
> on 2020/02/24 0:37, Darrick J. Wong wrote:
>> On Sun, Feb 23, 2020 at 10:22:10PM +0800, Eryu Guan wrote:
>>> Hi Darrick,
>>>
>>> On Thu, Feb 20, 2020 at 01:38:01PM +0800, Yang Xu wrote:
>>>> Currently, if we don't specify -l sunit or -l su option, mkfs.xfs
>>>> will get the stripe size from underlying device.
>>>>
>>>> It works file on most situations. But on some machine, the size of
>>>> underlying device greater than logbsize of mount options, it will
>>>> report error like "logbuf size must be greater than or equal to log
>>>> stripe size". We can specify -l su=4096 to meet this requirement and
>>>> case can still run normally.
>>>>
>>>> Also, from xfs manpage, version 2 also supports 16k log buf size for
>>>> mount option and case passed(only generic/054,055 used this api) on
>>>> my machine. So delete 32k and 64k with different sunit to be 
>>>> consistented
>>
>> I don't understand why there's a need to be consistent, which means I
>> don't understand why we'd "delete 32k and 64k with different sunit".
>> ext4 tests should not be invoking _xfs_log_config()
> 
> Sorry for the confused message.
> _get_log_configs{
> xfs)
>          _xfs_log_config
>          ;;
>      f2fs)
>          _f2fs_log_config
>          ;;
>      ext4)
>          _ext4_log_config
> 
> }
> xfs ,f2fs, and ext4 all have test1~test10.
> So for a generic case, it should be consistent.
> And only generic/054 055 uses _get_log_configs api and no test case
> used _xfs_log_config api. I think we should keep this(test1~test10, 
> 054.out only 10 times mkfs and mount output)
>>
>>>> with ext4 test num(10) and we can test all logbuf size.
>>
>> Hm?  ext4/010 is an inode bitmap fuzz test.
>>
> No. I mean _ext4_log_config has test1~test10.
> 
> Best Regards
> Yang Xu
>>>>
>>>> Signed-off-by: Yang Xu <xuyang2018.jy@cn.fujitsu.com>
>>>
>>> Would you like to review this v2 patch again? I feel more confident if
>>> xfs maintainer could ack it :)
>>>
>>> Thanks,
>>> Eryu
>>>
>>>> ---
>>>>   common/log | 12 ++++++------
>>>>   1 file changed, 6 insertions(+), 6 deletions(-)
>>>>
>>>> diff --git a/common/log b/common/log
>>>> index c7921f50..9b5a2f6d 100644
>>>> --- a/common/log
>>>> +++ b/common/log
>>>> @@ -546,15 +546,15 @@ _xfs_log_config()
>>>>   {
>>>>       echo "# mkfs-opt             mount-opt"
>>>>       echo "# ------------------------------"
>>>> -    echo "  version=2            logbsize=32k"
>>>> +    echo "  version=2,su=4096    logbsize=16k"
>>
>> The straight substitutions look fine to me, but then...
>>
>>>> +    echo "  version=2,su=16k     logbsize=16k"
>>
>> ...I can't tell why we're adding this extra case here, or why this
>> has to be here and not in a separate patch justifying the addition?
For 16k logbsize, I don't know much about it. it may make no sense? So
at the beginning, xfstests doesn't plan to test it. If so, I will remove 
this adding.

>>
>>>>       echo "  version=2,su=4096    logbsize=32k"
>>>> -    echo "  version=2,su=32768   logbsize=32k"
>>>> -    echo "  version=2,su=32768   logbsize=64k"
>>
>> ...I also don't think it's a good idea to reduce test coverage, and
>> definitely not buried in something that sounds like a fix patch.

OK. how about this change , as below(only specifying su=4096 for old 
options without su, logbusize=32K has su=4096, so add su=16k for it):
diff --git a/common/log b/common/log
index c7921f50..991c8e8f 100644
--- a/common/log
+++ b/common/log
@@ -546,15 +546,15 @@ _xfs_log_config()
  {
      echo "# mkfs-opt             mount-opt"
      echo "# ------------------------------"
-    echo "  version=2            logbsize=32k"
      echo "  version=2,su=4096    logbsize=32k"
+    echo "  version=2,su=16384   logbsize=32k"
      echo "  version=2,su=32768   logbsize=32k"
-    echo "  version=2,su=32768   logbsize=64k"
-    echo "  version=2            logbsize=64k"
+    echo "  version=2,su=4096    logbsize=64k"
+    echo "  version=2,su=32768   logbsize=64k"
      echo "  version=2,su=64k     logbsize=64k"
-    echo "  version=2            logbsize=128k"
+    echo "  version=2,su=4096    logbsize=128k"
      echo "  version=2,su=128k    logbsize=128k"
-    echo "  version=2            logbsize=256k"
+    echo "  version=2,su=4096    logbsize=256k"
      echo "  version=2,su=256k    logbsize=256k"

Best Regards
Yang Xu
>>
>> --D
>>
>>>> -    echo "  version=2            logbsize=64k"
>>>> +    echo "  version=2,su=32k     logbsize=32k"
>>>> +    echo "  version=2,su=4096    logbsize=64k"
>>>>       echo "  version=2,su=64k     logbsize=64k"
>>>> -    echo "  version=2            logbsize=128k"
>>>> +    echo "  version=2,su=4096    logbsize=128k"
>>>>       echo "  version=2,su=128k    logbsize=128k"
>>>> -    echo "  version=2            logbsize=256k"
>>>> +    echo "  version=2,su=4096    logbsize=256k"
>>>>       echo "  version=2,su=256k    logbsize=256k"
>>>>   }
>>>> -- 
>>>> 2.18.0
>>>>
>>>>
>>>>
>>
>>
> 
> 



      reply	other threads:[~2020-02-24  2:14 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-02-18  7:24 [PATCH v1 1/2] generic/054,055: add _require_v2log Yang Xu
2020-02-18  7:24 ` [PATCH v1 2/2] common/log: add 16k logbuf size for xfs_log_config Yang Xu
2020-02-19  5:06 ` [PATCH v1 1/2] generic/054,055: add _require_v2log Darrick J. Wong
2020-02-19  8:48   ` Yang Xu
2020-02-19 22:11     ` Darrick J. Wong
2020-02-20  2:01       ` Yang Xu
2020-02-20  2:17         ` Darrick J. Wong
2020-02-20  3:10           ` Yang Xu
2020-02-20  5:38           ` [PATCH v2] common/log: add -l su option in _xfs_log_config Yang Xu
2020-02-23 14:22             ` Eryu Guan
2020-02-23 16:37               ` Darrick J. Wong
2020-02-24  1:51                 ` Yang Xu
2020-02-24  2:14                   ` Yang Xu [this message]

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=484e9bf8-f659-c752-f232-11ee76a353a1@cn.fujitsu.com \
    --to=xuyang2018.jy@cn.fujitsu.com \
    --cc=darrick.wong@oracle.com \
    --cc=fstests@vger.kernel.org \
    --cc=guaneryu@gmail.com \
    /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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).