From: Yu Kuai <yukuai3@huawei.com>

After commit 4d27e927344a ("md: don't quiesce in mddev_suspend()"),
'conf->barrier' will be leaked in the case that raid10 takeover raid0:

level_store
 pers->takeover -> raid10_takeover
  raid10_takeover_raid0
   WRITE_ONCE(conf->barrier, 1)

mddev_suspend
// still raid0
mddev->pers = pers
// switch to raid10
mddev_resume
// resume without suspend

After the above commit, mddev_resume() will not decrease 'conf->barrier'
that is set in raid10_takeover_raid0().

Fix this problem by remove the setting of 'conf->barrier' from
raid10_takeover_raid0().

By the way, this problem is found while I'm trying to make
mddev_suspend/resume() to be independent from raid personalities. rai10
is the only personality to use reference count in the quiesce() callback
and this problem is only related to raid10.

Fixes: 4d27e927344a ("md: don't quiesce in mddev_suspend()")
Signed-off-by: Yu Kuai <yukuai3@huawei.com>
---
 drivers/md/raid10.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/drivers/md/raid10.c b/drivers/md/raid10.c
index 16aa9d735880..7704a4c7f469 100644
--- a/drivers/md/raid10.c
+++ b/drivers/md/raid10.c
@@ -4417,7 +4417,6 @@ static void *raid10_takeover_raid0(struct mddev *mddev, sector_t size, int devs)
 				rdev->new_raid_disk = rdev->raid_disk * 2;
 				rdev->sectors = size;
 			}
-		WRITE_ONCE(conf->barrier, 1);
 	}
 
 	return conf;
-- 
2.39.2

Dear Yu,


Thank you for your patch. Some minor style comments for the commit 
message, you can also ignore.

Am 31.07.23 um 04:28 schrieb Yu Kuai:
> From: Yu Kuai <yukuai3@huawei.com>
> 
> After commit 4d27e927344a ("md: don't quiesce in mddev_suspend()"),
> 'conf->barrier' will be leaked in the case that raid10 takeover raid0:
> 
> level_store
>   pers->takeover -> raid10_takeover
>    raid10_takeover_raid0
>     WRITE_ONCE(conf->barrier, 1)
> 
> mddev_suspend
> // still raid0
> mddev->pers = pers
> // switch to raid10
> mddev_resume
> // resume without suspend
> 
> After the above commit, mddev_resume() will not decrease 'conf->barrier'
> that is set in raid10_takeover_raid0().
> 
> Fix this problem by remove the setting of 'conf->barrier' from

removing

or

… by not setting `conf->barrier` in …

> raid10_takeover_raid0().
> 
> By the way, this problem is found while I'm trying to make
> mddev_suspend/resume() to be independent from raid personalities. rai10

rai*d*

> is the only personality to use reference count in the quiesce() callback
> and this problem is only related to raid10.
> 
> Fixes: 4d27e927344a ("md: don't quiesce in mddev_suspend()")
> Signed-off-by: Yu Kuai <yukuai3@huawei.com>
> ---
>   drivers/md/raid10.c | 1 -
>   1 file changed, 1 deletion(-)
> 
> diff --git a/drivers/md/raid10.c b/drivers/md/raid10.c
> index 16aa9d735880..7704a4c7f469 100644
> --- a/drivers/md/raid10.c
> +++ b/drivers/md/raid10.c
> @@ -4417,7 +4417,6 @@ static void *raid10_takeover_raid0(struct mddev *mddev, sector_t size, int devs)
>   				rdev->new_raid_disk = rdev->raid_disk * 2;
>   				rdev->sectors = size;
>   			}
> -		WRITE_ONCE(conf->barrier, 1);
>   	}
>   
>   	return conf;

Reviewed-by: Paul Menzel <pmenzel@molgen.mpg.de>


Kind regards,

Paul

On Mon, Jul 31, 2023 at 9:50 AM Paul Menzel <pmenzel@molgen.mpg.de> wrote:
[...]
> >       return conf;
>
> Reviewed-by: Paul Menzel <pmenzel@molgen.mpg.de>
>

Applied to md-next after incorporating Paul's recommendations.

Thanks,
Song

>
> Kind regards,
>
> Paul