Re: [PATCH RESEND v2] btrfs: fix devid 0 without a replace item by failing the mount

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

>> diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
>> index 1b2742da5d4a..bb6f067f2fb9 100644
>> --- a/fs/btrfs/volumes.c
>> +++ b/fs/btrfs/volumes.c
>> @@ -1056,22 +1056,13 @@ static void __btrfs_free_extra_devids(struct btrfs_fs_devices *fs_devices,
>>   			continue;
>>   		}
>>   
>> -		if (device->devid == BTRFS_DEV_REPLACE_DEVID) {
>> -			/*
>> -			 * In the first step, keep the device which has
>> -			 * the correct fsid and the devid that is used
>> -			 * for the dev_replace procedure.
>> -			 * In the second step, the dev_replace state is
>> -			 * read from the device tree and it is known
>> -			 * whether the procedure is really active or
>> -			 * not, which means whether this device is
>> -			 * used or whether it should be removed.
>> -			 */
>> -			if (step == 0 || test_bit(BTRFS_DEV_STATE_REPLACE_TGT,
> 
> This removes the use of step parameter so it can be removed from the
> call chain too (separate patch).

  Will see to it.

> Patch added to misc-next, thanks.

Thanks.