Re: [f2fs-dev] [PATCH] f2fs: avoid f2fs_map_blocks() for consecutive holes in readpages

[Chao Yu <chao@kernel.org>]

On 1/19/2026 3:43 PM, Nanzhe Zhao wrote:
> Hi Chao:
> At 2026-01-16 16:52:02, "Chao Yu" <chao@kernel.org> wrote:
>>>
>>> Do we have plans to also support reducing f2fs_map_blocks() calls for
>>> consectives holes in logical file postion with dnode have already been
>>> allocated in buffered large folio read?
>>> Such as consective NULL_ADDR or NEW_ADDR?
>>
>> Nanzhe,
>>
>> We have supported that for large folio read w/ this patch?
>>
>> Thanks,
>>
> 
> Sorry, I'm a bit confused.
> In the condition of F2FS_MAP_BLOCK_DEFAULT, the default: case will only
> set map->m_next_pgofs to pgofs + 1 then sync out. When we enter
> next iteration and the index advanced, currrent index now turns to pgofs + 1
> and index < next_pgofs become false.In consequence, we won't reduce
> f2fs_map_blocks() calls for hole with dnode allocated.
> 
> Also, for NEW_ADDR, the default: case will directly go to sync out and bypass
> map_is_mergeable, so it will also not reduce f2fs_map_blocks calls.
> 
> Or am I missing something?

I guess f2fs_map_no_dnode() will update map->m_next_pgofs to pgofs of next
potential valid dnode.

Thanks,

> 
> Thanks,
> Nanzhe Zhao