From mboxrd@z Thu Jan  1 00:00:00 1970
From: Xiao Yang <yangx.jy@cn.fujitsu.com>
Date: Sun, 24 Jun 2018 13:25:55 +0800
Subject: [LTP] [PATCH] commands/mkswap01.sh: Revert the original retry
	loops
In-Reply-To: <CAEemH2c4tC1Lk5bu2Tym9N5ej2yuqR15U+ApO-+vZb8XsKm-Jg@mail.gmail.com>
References: <1529551625-366-1-git-send-email-yangx.jy@cn.fujitsu.com>
 <CAEemH2c4tC1Lk5bu2Tym9N5ej2yuqR15U+ApO-+vZb8XsKm-Jg@mail.gmail.com>
Message-ID: <5B2F2B63.3010401@cn.fujitsu.com>
List-Id: <ltp.lists.linux.it>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
To: ltp@lists.linux.it

On 2018/06/22 18:20, Li Wang wrote:
> Hi Xiao,
>
> Xiao Yang<yangx.jy@cn.fujitsu.com>  wrote:
>> Before kernel commit c3473c6(e.g. RHEL6), mkswap -L or mkswap -U for loop
>> devices can not generate corresponding "/dev/disk/by-*" files and gets TBROK.
>> It is just a way to check the success of mkswap rather than a bug, so the test
>> should not get TBROK when "/dev/disk/by-*" files are not generated but swapon
>> enable loop devices for swapping sucessfully.
> Thanks for reporting this.
>
>> Signed-off-by: Xiao Yang<yangx.jy@cn.fujitsu.com>
>> ---
>>   testcases/commands/mkswap/mkswap01.sh | 19 +++++++++++++++----
>>   1 file changed, 15 insertions(+), 4 deletions(-)
>>
>> diff --git a/testcases/commands/mkswap/mkswap01.sh b/testcases/commands/mkswap/mkswap01.sh
>> index 5ead4cd..1c01c86 100755
>> --- a/testcases/commands/mkswap/mkswap01.sh
>> +++ b/testcases/commands/mkswap/mkswap01.sh
>> @@ -37,14 +37,25 @@ setup()
>>          DEVICE_SIZE=$((($real_size/$PAGE_SIZE * $PAGE_SIZE)/1024))
>>   }
>>
>> -check_for_file()
>> +wait_for_file()
>>   {
>>          local path="$1"
>> +       local retries=10
>>
>> -       if [ -z "$path" -o -e "$path" ]; then
>> +       if [ -z "$path" ]; then
>>                  return
>>          fi
>> -       return 1
>> +
>> +       while [ $retries -gt 0 ]; do
>> +               if [ -e "$path" ]; then
>> +                       return
>> +               fi
>> +               tst_res TINFO "Waiting for $path to appear"
>> +               retries=$((retries - 1))
>> +               tst_sleep 10ms
>> +       done
>> +
>> +       tst_res TINFO "The file $path haven't appeared"
>>   }
>>
>>   mkswap_verify()
>> @@ -64,7 +75,7 @@ mkswap_verify()
>>                  local pagesize=$PAGE_SIZE
>>          fi
>>
>> -       TST_RETRY_FUNC "check_for_file $dev_file" 0
>> +       wait_for_file "$dev_file"
>>
>>          swapon $swapfile 2>/dev/null
> To revert this patch is also not wise for old kernels (without commit
> c3473c63542), because the kernel has to finish this looping to wait
> for a disk label that will never appear.
>
> So, what about just wait 100ms for old kernels directly?
Hi Li,

OK, it is reasonabel to me, and i will send v2 patch.

Thanks,
Xiao Yang
> --- a/testcases/commands/mkswap/mkswap01.sh
> +++ b/testcases/commands/mkswap/mkswap01.sh
> @@ -64,7 +64,11 @@ mkswap_verify()
>                  local pagesize=$PAGE_SIZE
>          fi
>
> -       TST_RETRY_FUNC "check_for_file $dev_file" 0
> +       if tst_kvcmp -lt '2.6.35'; then
> +               tst_resm TINFO "Waiting for $dev_file to appear"
> +               tst_sleep 100ms
> +       else
> +               TST_RETRY_FUNC "check_for_file $dev_file" 0
>
>          swapon $swapfile 2>/dev/null
>
>