2020-02-26 09:57:57

by Ritesh Harjani

Subject: [PATCHv3 1/6] ext4: Add IOMAP_F_MERGED for non-extent based mapping

IOMAP_F_MERGED needs to be set in case of non-extent based mapping.
This is needed in later patches for conversion of ext4_fiemap to use iomap.

Signed-off-by: Ritesh Harjani <[email protected]>
---
fs/ext4/inode.c | 4 ++++
1 file changed, 4 insertions(+)

diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index d035acab5b2a..3b4230cf0bc2 100644
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -3335,6 +3335,10 @@ static void ext4_set_iomap(struct inode *inode, struct iomap *iomap,
iomap->offset = (u64) map->m_lblk << blkbits;
iomap->length = (u64) map->m_len << blkbits;

+ if ((map->m_flags & (EXT4_MAP_UNWRITTEN | EXT4_MAP_MAPPED)) &&
+ !ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))
+ iomap->flags |= IOMAP_F_MERGED;
+
/*
* Flags passed to ext4_map_blocks() for direct I/O writes can result
* in m_flags having both EXT4_MAP_MAPPED and EXT4_MAP_UNWRITTEN bits
--
2.21.0

2020-02-26 12:32:23

by Jan Kara

Subject: Re: [PATCHv3 1/6] ext4: Add IOMAP_F_MERGED for non-extent based mapping

On Wed 26-02-20 15:27:03, Ritesh Harjani wrote:
> IOMAP_F_MERGED needs to be set in case of non-extent based mapping.
> This is needed in later patches for conversion of ext4_fiemap to use iomap.
>
> Signed-off-by: Ritesh Harjani <[email protected]>
> ---
> fs/ext4/inode.c | 4 ++++
> 1 file changed, 4 insertions(+)
>
> diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
> index d035acab5b2a..3b4230cf0bc2 100644
> --- a/fs/ext4/inode.c
> +++ b/fs/ext4/inode.c
> @@ -3335,6 +3335,10 @@ static void ext4_set_iomap(struct inode *inode, struct iomap *iomap,
> iomap->offset = (u64) map->m_lblk << blkbits;
> iomap->length = (u64) map->m_len << blkbits;
>
> + if ((map->m_flags & (EXT4_MAP_UNWRITTEN | EXT4_MAP_MAPPED)) &&
> + !ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))
^^ we indent the second line of condition either like:

if ((map->m_flags & (EXT4_MAP_UNWRITTEN | EXT4_MAP_MAPPED)) &&
!ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))

or like:

if ((map->m_flags & (EXT4_MAP_UNWRITTEN | EXT4_MAP_MAPPED)) &&
!ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))

But on at the same level as following code block because that's highly
confusing at times...

Also EXT4_MAP_UNWRITTEN is never set for indirect block mapped files (that
on disk format does not support unwritten extents).

Honza

> + iomap->flags |= IOMAP_F_MERGED;
> +
> /*
> * Flags passed to ext4_map_blocks() for direct I/O writes can result
> * in m_flags having both EXT4_MAP_MAPPED and EXT4_MAP_UNWRITTEN bits
> --
> 2.21.0
>
--
Jan Kara <[email protected]>
SUSE Labs, CR

2020-02-26 12:35:56

by Ritesh Harjani

Subject: Re: [PATCHv3 1/6] ext4: Add IOMAP_F_MERGED for non-extent based mapping

On 2/26/20 5:56 PM, Jan Kara wrote:
> On Wed 26-02-20 15:27:03, Ritesh Harjani wrote:
>> IOMAP_F_MERGED needs to be set in case of non-extent based mapping.
>> This is needed in later patches for conversion of ext4_fiemap to use iomap.
>>
>> Signed-off-by: Ritesh Harjani <[email protected]>
>> ---
>> fs/ext4/inode.c | 4 ++++
>> 1 file changed, 4 insertions(+)
>>
>> diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
>> index d035acab5b2a..3b4230cf0bc2 100644
>> --- a/fs/ext4/inode.c
>> +++ b/fs/ext4/inode.c
>> @@ -3335,6 +3335,10 @@ static void ext4_set_iomap(struct inode *inode, struct iomap *iomap,
>> iomap->offset = (u64) map->m_lblk << blkbits;
>> iomap->length = (u64) map->m_len << blkbits;
>>
>> + if ((map->m_flags & (EXT4_MAP_UNWRITTEN | EXT4_MAP_MAPPED)) &&
>> + !ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))
> ^^ we indent the second line of condition either like:
>
> if ((map->m_flags & (EXT4_MAP_UNWRITTEN | EXT4_MAP_MAPPED)) &&
> !ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))

Yes, I too prefer it this way. Will fix it.

>
> or like:
>
> if ((map->m_flags & (EXT4_MAP_UNWRITTEN | EXT4_MAP_MAPPED)) &&
> !ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))
>
> But on at the same level as following code block because that's highly
> confusing at times...
>
> Also EXT4_MAP_UNWRITTEN is never set for indirect block mapped files (that
> on disk format does not support unwritten extents).

Aah yes. Let me fix it.

>
> Honza
>
>> + iomap->flags |= IOMAP_F_MERGED;
>> +
>> /*
>> * Flags passed to ext4_map_blocks() for direct I/O writes can result
>> * in m_flags having both EXT4_MAP_MAPPED and EXT4_MAP_UNWRITTEN bits
>> --
>> 2.21.0
>>