[Cluster-devel] [Ocfs2-devel] [PATCH 1/3] fs/buffer.c: add new api to allow eof writeback

[Junxiao Bi]

On 5/9/21 4:23 PM, Andrew Morton wrote:

> On Mon, 26 Apr 2021 15:05:50 -0700 Junxiao Bi <junxiao.bi at oracle.com> wrote:
>
>> When doing truncate/fallocate for some filesytem like ocfs2, it
>> will zero some pages that are out of inode size and then later
>> update the inode size, so it needs this api to writeback eof
>> pages.
> Seems reasonable.  But can we please update the
> __block_write_full_page_eof() comment?  It now uses the wrong function
> name and doesn't document the new `eof' argument.

Jan suggested using sb_issue_zeroout to zero eof pages in 
ocfs2_fallocate, that can

also fix the issue for ocfs2. For gfs2, i though it had the same issue, 
but i didn't get

a confirm from gfs2 maintainer, if gfs2 is ok, then maybe this new api 
is not necessary?

Thanks,

Junxiao.

>