Re: [RFC PATCH] fuse: invalidate page cache pages before direct write

[Bernd Schubert <bernd.schubert@fastmail.fm>]

On 6/8/23 09:17, Hao Xu wrote:
> Ping...
> 
> On 5/9/23 16:01, Hao Xu wrote:
>> From: Hao Xu <howeyxu@tencent.com>
>>
>> In FOPEN_DIRECT_IO, page cache may still be there for a file, direct
>> write should respect that and invalidate the corresponding pages so
>> that page cache readers don't get stale data. Another thing this patch
>> does is flush related pages to avoid its loss.
>>
>> Signed-off-by: Hao Xu <howeyxu@tencent.com>
>> ---
>>
>> Reference:
>> https://lore.kernel.org/linux-fsdevel/ee8380b3-683f-c526-5f10-1ce2ee6f79ad@linux.dev/#:~:text=I%20think%20this%20problem%20exists%20before%20this%20patchset
>>
>>   fs/fuse/file.c | 14 +++++++++++++-
>>   1 file changed, 13 insertions(+), 1 deletion(-)
>>
>> diff --git a/fs/fuse/file.c b/fs/fuse/file.c
>> index 89d97f6188e0..edc84c1dfc5c 100644
>> --- a/fs/fuse/file.c
>> +++ b/fs/fuse/file.c
>> @@ -1490,7 +1490,8 @@ ssize_t fuse_direct_io(struct fuse_io_priv *io, 
>> struct iov_iter *iter,
>>       int write = flags & FUSE_DIO_WRITE;
>>       int cuse = flags & FUSE_DIO_CUSE;
>>       struct file *file = io->iocb->ki_filp;
>> -    struct inode *inode = file->f_mapping->host;
>> +    struct address_space *mapping = file->f_mapping;
>> +    struct inode *inode = mapping->host;
>>       struct fuse_file *ff = file->private_data;
>>       struct fuse_conn *fc = ff->fm->fc;
>>       size_t nmax = write ? fc->max_write : fc->max_read;
>> @@ -1516,6 +1517,17 @@ ssize_t fuse_direct_io(struct fuse_io_priv *io, 
>> struct iov_iter *iter,
>>               inode_unlock(inode);
>>       }
>> +    res = filemap_write_and_wait_range(mapping, pos, pos + count - 1);
>> +    if (res)
>> +        return res;
>> +
>> +    if (write) {
>> +        if (invalidate_inode_pages2_range(mapping,
>> +                idx_from, idx_to)) {
>> +            return -ENOTBLK;
>> +        }
>> +    }
>> +
>>       io->should_dirty = !write && user_backed_iter(iter);
>>       while (count) {
>>           ssize_t nres;
> 

Is this part not working?

	if (!cuse && fuse_range_is_writeback(inode, idx_from, idx_to)) {
		if (!write)
			inode_lock(inode);
		fuse_sync_writes(inode);
		if (!write)
			inode_unlock(inode);
	}



Thanks,
Bernd