Re: [PATCH v6 2/6] fstests: check: fix unset seqres in run_section()

[Anand Jain <anand.jain@oracle.com>]

On 10/4/25 05:31, Anand Jain wrote:
> 
> 
> On 9/4/25 17:57, Zorro Lang wrote:
>> On Wed, Apr 09, 2025 at 03:43:14PM +0800, Anand Jain wrote:
>>> Ensure seqres is set early in run_section().
>>>
>>> Signed-off-by: Anand Jain <anand.jain@oracle.com>
>>> ---
>>>   check | 2 +-
>>>   1 file changed, 1 insertion(+), 1 deletion(-)
>>>
>>> diff --git a/check b/check
>>> index 32890470a020..16f695e9d75c 100755
>>> --- a/check
>>> +++ b/check
>>> @@ -804,6 +804,7 @@ function run_section()
>>>       seq="check.$$"
>>>       check="$RESULT_BASE/check"
>>> +    seqres="$check"
>>
>> The "seqres" even might be used earlier than that. If your rootfs is 
>> readonly,
>> you'll see that.

Zorro,

While testing, I saw that ".full" is created by _scratch_unmount(), 
called here:


  725 function run_section()
  ::
  815         if [ ! -z "$SCRATCH_DEV" ]; then
  816           _scratch_unmount 2> /dev/null
  817           # call the overridden mkfs - make sure the FS is built


So the fix in this patch is correct..

I also tried with a read-only rootfs—".full" wasn’t
there an error to create that file..

I’ll resend the patch soon.

Thanks. Anand


>>
> 
> Zorro,
> 
> Thanks a lot for the review and RVB!
> 
> I’ll take care of this patch 2/6 in a separate patchset.
> Meanwhile, could you help merge the rest of the sysfs patches,
> except for patch 2/6? I don't want the seqres issue to block
> the rest of the sysfs patches.
> 
> Thanks, Anand
> 
>> Thanks,
>> Zorro
>>
>>>       # don't leave old full output behind on a clean run
>>>       rm -f $check.full
>>> @@ -849,7 +850,6 @@ function run_section()
>>>         fi
>>>       fi
>>> -    seqres="$check"
>>>       _check_test_fs
>>>       loop_status=()    # track rerun-on-failure state
>>> -- 
>>> 2.47.0
>>>
>>
>