On 2024/1/3 19:31, Jan Kara wrote:
> On Tue 02-01-24 20:38:58, Zhang Yi wrote:
>> From: Zhang Yi <yi.zhang@xxxxxxxxxx>
>>
>> Add a new map flag EXT4_MAP_DELAYED to indicate the mapping range is a
>> delayed allocated only (not unwritten) one, and making
>> ext4_map_blocks() can distinguish it, no longer mixing it with holes.
>>
>> Signed-off-by: Zhang Yi <yi.zhang@xxxxxxxxxx>
>
> One small comment below.
>
>> ---
>> fs/ext4/ext4.h | 4 +++-
>> fs/ext4/extents.c | 5 +++--
>> fs/ext4/inode.c | 2 ++
>> 3 files changed, 8 insertions(+), 3 deletions(-)
>>
>> diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
>> index a5d784872303..55195909d32f 100644
>> --- a/fs/ext4/ext4.h
>> +++ b/fs/ext4/ext4.h
>> @@ -252,8 +252,10 @@ struct ext4_allocation_request {
>> #define EXT4_MAP_MAPPED BIT(BH_Mapped)
>> #define EXT4_MAP_UNWRITTEN BIT(BH_Unwritten)
>> #define EXT4_MAP_BOUNDARY BIT(BH_Boundary)
>> +#define EXT4_MAP_DELAYED BIT(BH_Delay)
>> #define EXT4_MAP_FLAGS (EXT4_MAP_NEW | EXT4_MAP_MAPPED |\
>> - EXT4_MAP_UNWRITTEN | EXT4_MAP_BOUNDARY)
>> + EXT4_MAP_UNWRITTEN | EXT4_MAP_BOUNDARY |\
>> + EXT4_MAP_DELAYED)
>>
>> struct ext4_map_blocks {
>> ext4_fsblk_t m_pblk;
>> diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c
>> index 0892d0568013..fc69f13cf510 100644
>> --- a/fs/ext4/extents.c
>> +++ b/fs/ext4/extents.c
>> @@ -4073,9 +4073,10 @@ static void ext4_ext_determine_hole(struct inode *inode,
>> } else if (in_range(map->m_lblk, es.es_lblk, es.es_len)) {
>> /*
>> * Straddle the beginning of the queried range, it's no
>> - * longer a hole, adjust the length to the delayed extent's
>> - * after map->m_lblk.
>> + * longer a hole, mark it is a delalloc and adjust the
>> + * length to the delayed extent's after map->m_lblk.
>> */
>> + map->m_flags |= EXT4_MAP_DELAYED;
>
> I wouldn't set delalloc bit here. If there's delalloc extent containing
> lblk now, it means the caller of ext4_map_blocks() is not holding i_rwsem
> (otherwise we would have found already in ext4_map_blocks()) and thus
> delalloc info is unreliable anyway. So I wouldn't bother. But it's worth a
> comment here like:
>
> /*
> * There's delalloc extent containing lblk. It must have
> * been added after ext4_map_blocks() checked the extent
> * status tree so we are not holding i_rwsem and delalloc
> * info is only stabilized by i_data_sem we are going to
> * release soon. Don't modify the extent status tree and
> * report extent as a hole.
> */
>
Yeah, the delalloc info is still unreliable. Thanks for the advice, I
will revise it in my next iteration along with your advice in patch 03.
Thanks,
Yi.
>
>> len = es.es_lblk + es.es_len - map->m_lblk;
>> goto out;
>> } else {
>> diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
>> index 1b5e6409f958..c141bf6d8db2 100644
>> --- a/fs/ext4/inode.c
>> +++ b/fs/ext4/inode.c
>> @@ -515,6 +515,8 @@ int ext4_map_blocks(handle_t *handle, struct inode *inode,
>> map->m_len = retval;
>> } else if (ext4_es_is_delayed(&es) || ext4_es_is_hole(&es)) {
>> map->m_pblk = 0;
>> + map->m_flags |= ext4_es_is_delayed(&es) ?
>> + EXT4_MAP_DELAYED : 0;
>> retval = es.es_len - (map->m_lblk - es.es_lblk);
>> if (retval > map->m_len)
>> retval = map->m_len;
>> --
>> 2.39.2
>>
