Re: [PATCH] btrfs/154: test for device dynamic rescan

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

On 10/21/2018 05:31 PM, Filipe Manana wrote:
> On Sun, Oct 21, 2018 at 10:20 AM Nikolay Borisov <nborisov@suse.com> wrote:
>>
>>
>>
>> On 21.10.2018 10:16, Filipe Manana wrote:
>>> On Mon, Nov 13, 2017 at 2:26 AM Anand Jain <anand.jain@oracle.com> wrote:
>>>>
>>>> Make sure missing device is included in the alloc list when it is
>>>> scanned on a mounted FS.
>>>>
>>>> This test case needs btrfs kernel patch which is in the ML
>>>>    [PATCH] btrfs: handle dynamically reappearing missing device
>>>> Without the kernel patch, the test will run, but reports as
>>>> failed, as the device scanned won't appear in the alloc_list.
>>>
>>> So that patch was never merged, at least not with that subject.
>>> What happened?

  The contending patch at V5 (with fixed comments as received) was
  working fine at the then latest kernel version, now its out dated,
  and I am not sure where to invest my time, it will be better to
  have David's comment, if he has any concern on the idea/approach
  in general. I am ok to revive the patch.

>> https://git.kernel.org/pub/scm/fs/xfs/xfstests-dev.git/tree/tests/btrfs/154
>>
>> In my testing of misc-next this test has been failing for me.

  Don't bother about that patch I am sure its not good on the current
  kernel.

>  recent hole punch
> test, that doesn't have a corresponding kernel fix.

  Hole punching kernel patch, fixes a corner case, I tried but failed.
  Until my current priority list recedes I won't be able to look into
  it again. If you would like to pick this up pls go ahead. I think it
  will easy for you.

  Lastly, I think its a good idea to record the current known limitation
  as a test cases, even if it fails, in that way it gets documented.

Thanks, Anand