Re: [RFC PATCH 6/8] fat: add iomap buffered I/O support

From: Matheus Alves de Almeida

Date: Wed Nov 11 2026 - 23:36:43 EST


On Mon, Nov 09, 2026 at 12:03:00AM +0900, OGAWA Hirofumi wrote:
>Matheus Alves de Almeida <matheus.aalmeida@xxxxxxxxxxxx> writes:
>
>> Use iomap for buffered I/O, replacing buffer_heads. Remove
>> fat_get_block(), fat_write_begin(), fat_write_failed() as the iomap path
>> makes them unnecessary.
>>
>> Signed-off-by: Matheus Alves de Almeida <matheus.aalmeida@xxxxxxxxxxxx>
>> ---
>> fs/fat/fat.h | 4 +-
>> fs/fat/file.c | 72 +++++++++++++++++++----
>> fs/fat/inode.c | 157 +++++--------------------------------------------
>> fs/fat/iomap.c | 84 ++++++++++++++++++++++++++
>> 4 files changed, 164 insertions(+), 153 deletions(-)
>>
>
>[...]
>
>> diff --git a/fs/fat/file.c b/fs/fat/file.c
>> index 8689cbbdad13..723b671d5b4b 100644
>> --- a/fs/fat/file.c
>> +++ b/fs/fat/file.c
>
>[...]
>
>> +static int fat_cont_expand(struct inode *inode, loff_t new_size);
>> +
>> +static ssize_t fat_file_write_iter_buffered(struct kiocb *iocb,
>> + struct iov_iter *iter)
>> +{
>> + struct inode *inode = file_inode(iocb->ki_filp);
>> + struct msdos_sb_info *sbi = MSDOS_SB(inode->i_sb);
>> + loff_t offset, old_size, old_alloc, keep_alloc;
>> + ssize_t ret;
>> +
>> + inode_lock(inode);
>> + ret = generic_write_checks(iocb, iter);
>> + if (ret <= 0)
>> + goto out;
>> +
>> + ret = file_modified(iocb->ki_filp);
>> + if (ret)
>> + goto out;
>> +
>> + offset = iocb->ki_pos;
>> + old_size = i_size_read(inode);
>> + if (offset > old_size) {
>> + /* We need to fill this hole */
>> + ret = fat_cont_expand(inode, offset);
>> + if (ret)
>> + goto out;
>> + }
>> +
>> + inode_dio_wait(inode);
>> + old_alloc = inode_get_bytes(inode);
>> + ret = iomap_file_buffered_write(iocb, iter, &fat_iomap_write_ops,
>> + NULL, NULL);
>> + keep_alloc = max(old_alloc, round_up(i_size_read(inode),
>> + sbi->cluster_size));
>> + if (inode_get_bytes(inode) > keep_alloc)
>> + fat_truncate_blocks(inode, keep_alloc);
>
>What is doing by this fat_truncate_blocks?
>

It's trimming the clusters past EOF and keeping the fallocate KEEP_SIZE
preallocation. I can add a comment stating this.

>> + if (ret > 0)
>> + ret = generic_write_sync(iocb, ret);
>
>Maybe we want to do this outside inode_unlock()?
>

True, good catch.

>> +out:
>> + inode_unlock(inode);
>> + return ret;
>> +}
>> +

Thanks,
Matheus