Re: [f2fs-dev] [PATCH 2/8] f2fs: convert f2fs_write_end() to use folio

[Chao Yu <chao@kernel.org>]

On 2024/8/20 11:04, Li Zetao wrote:
> Hi,
> 
> 在 2024/8/19 9:20, Chao Yu 写道:
>> Convert to use folio, so that we can get rid of 'page->index' to
>> prepare for removal of 'index' field in structure page [1].
>>
>> [1] https://lore.kernel.org/all/Zp8fgUSIBGQ1TN0D@casper.infradead.org/
>>
>> Cc: Matthew Wilcox <willy@infradead.org>
>> Signed-off-by: Chao Yu <chao@kernel.org>
>> ---
>>   fs/f2fs/data.c | 19 ++++++++++---------
>>   1 file changed, 10 insertions(+), 9 deletions(-)
>>
>> diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c
>> index 9e1ade13060c..e114d738b6b4 100644
>> --- a/fs/f2fs/data.c
>> +++ b/fs/f2fs/data.c
>> @@ -3378,7 +3378,7 @@ void f2fs_write_failed(struct inode *inode, loff_t to)
>>   }
>>   static int prepare_write_begin(struct f2fs_sb_info *sbi,
>> -            struct page *page, loff_t pos, unsigned len,
>> +            struct folio *folio, loff_t pos, unsigned int len,
>>               block_t *blk_addr, bool *node_changed)
>>   {
>>       struct inode *inode = folio->mapping->host;
>> @@ -3425,7 +3425,7 @@ static int prepare_write_begin(struct f2fs_sb_info *sbi,
>>                   set_page_private_inline(ipage);
>>               goto out;
>>           }
>> -        err = f2fs_convert_inline_page(&dn, page);
>> +        err = f2fs_convert_inline_page(&dn, folio_page(folio, 0));
>>           if (err || dn.data_blkaddr != NULL_ADDR)
>>               goto out;
>>       }
>> @@ -3629,7 +3629,7 @@ static int f2fs_write_begin(struct file *file, struct address_space *mapping,
>>       folio = page_folio(page);
>>       if (f2fs_is_atomic_file(inode))
>> -        err = prepare_atomic_write_begin(sbi, folio, pos, len
>> +        err = prepare_atomic_write_begin(sbi, folio, pos, len,

Thanks for your reminder, I removed comma after @len in patch 1/8 by
mistake, let me fix it.

Thanks,

> This has been changed in the first patch, it is redundant here.
>>                       &blkaddr, &need_balance, &use_cow);
>>       else
>>           err = prepare_write_begin(sbi, folio, pos, len,
>> @@ -3698,7 +3698,8 @@ static int f2fs_write_end(struct file *file,
>>               loff_t pos, unsigned len, unsigned copied,
>>               struct page *page, void *fsdata)
>>   {
>> -    struct inode *inode = page->mapping->host;
>> +    struct folio *folio = page_folio(page);
>> +    struct inode *inode = folio->mapping->host;
>>       trace_f2fs_write_end(inode, pos, len, copied);
>> @@ -3707,17 +3708,17 @@ static int f2fs_write_end(struct file *file,
>>        * should be PAGE_SIZE. Otherwise, we treat it with zero copied and
>>        * let generic_perform_write() try to copy data again through copied=0.
>>        */
>> -    if (!PageUptodate(page)) {
>> +    if (!folio_test_uptodate(folio)) {
>>           if (unlikely(copied != len))
>>               copied = 0;
>>           else
>> -            SetPageUptodate(page);
>> +            folio_mark_uptodate(folio);
>>       }
>>   #ifdef CONFIG_F2FS_FS_COMPRESSION
>>       /* overwrite compressed file */
>>       if (f2fs_compressed_file(inode) && fsdata) {
>> -        f2fs_compress_write_end(inode, fsdata, page->index, copied);
>> +        f2fs_compress_write_end(inode, fsdata, folio->index, copied);
>>           f2fs_update_time(F2FS_I_SB(inode), REQ_TIME);
>>           if (pos + copied > i_size_read(inode) &&
>> @@ -3730,10 +3731,10 @@ static int f2fs_write_end(struct file *file,
>>       if (!copied)
>>           goto unlock_out;
>> -    set_page_dirty(page);
>> +    folio_mark_dirty(folio);
>>       if (f2fs_is_atomic_file(inode))
>> -        set_page_private_atomic(page);
>> +        set_page_private_atomic(folio_page(folio, 0));
>>       if (pos + copied > i_size_read(inode) &&
>>           !f2fs_verity_in_progress(inode)) {



_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel