On 2018/10/12 下午2:52, Nikolay Borisov wrote:
>
>
> On 12.10.2018 09:37, Qu Wenruo wrote:
>> The member log_root_transid is never used.
>> It's always kept untouched even when updating log tree root.
>>
>> And populating it without introducing new incompat flags could easily
>> cause back-compatibility problems.
>> So just mark it unused.
>>
>> Signed-off-by: Qu Wenruo <wqu@xxxxxxxx>
>> ---
>> changelog:
>> v2:
>> Remove the redundant comment.
>> ---
>> fs/btrfs/ctree.h | 5 +----
>> 1 file changed, 1 insertion(+), 4 deletions(-)
>>
>> diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h
>> index 53af9f5253f4..7adf5f4dcda4 100644
>> --- a/fs/btrfs/ctree.h
>> +++ b/fs/btrfs/ctree.h
>> @@ -214,8 +214,7 @@ struct btrfs_super_block {
>> __le64 chunk_root;
>> __le64 log_root;
>>
>> - /* this will help find the new super based on the log root */
>> - __le64 log_root_transid;
>> + __le64 __unused_log_root_transid;
>
> why do you keep insisting on having "log_root_transid" in the name of
> this member? It was never used on-disk so for all intents and purposes
> log_root_transid doesn't mean anything? No need to resend, I'm sure
> David will change it on commit, nut I'm just curious.
We have a similar case, leaf_size.
Although it differs from this case and it's still used, I prefer to have
something showing that we at least planned to use this member.
Just in case one day we tries to use it again.
Thanks,
Qu
>
>
>> __le64 total_bytes;
>> __le64 bytes_used;
>> __le64 root_dir_objectid;
>> @@ -2317,8 +2316,6 @@ BTRFS_SETGET_STACK_FUNCS(super_chunk_root_level, struct btrfs_super_block,
>> chunk_root_level, 8);
>> BTRFS_SETGET_STACK_FUNCS(super_log_root, struct btrfs_super_block,
>> log_root, 64);
>> -BTRFS_SETGET_STACK_FUNCS(super_log_root_transid, struct btrfs_super_block,
>> - log_root_transid, 64);
>> BTRFS_SETGET_STACK_FUNCS(super_log_root_level, struct btrfs_super_block,
>> log_root_level, 8);
>> BTRFS_SETGET_STACK_FUNCS(super_total_bytes, struct btrfs_super_block,
>>
