Hi Al,
On Sun, Sep 01, 2019 at 04:05:14AM +0100, Al Viro wrote:
> On Sun, Sep 01, 2019 at 09:37:19AM +0800, Gao Xiang wrote:
>
> > fs/orangefs/file.c
> > 19 static int flush_racache(struct inode *inode)
>
> Just why the hell would _that_ one be a problem? It's static in
> file; it can't pollute the namespace even if linked into the
> kernel.
>
> Folks, let's keep at least some degree of sanity - this is sinking
> to the level of certain killfile denizens...
Thanks for your kind reply. I think in the same way.
And Christoph did many great suggestions for erofs, thanks him
for erofs, and I'm already fixed most of them, and some
suggestions I have no idea to do....
1) add "erofs_" to all functions [1] [2];
2) avoid sb_bread and use read_mapping_page, actually
read_mapping_page will call block_read_full_page and
buffer_heads still there;
and I don't know what erofs "rapidly turning" means, all great
suggestions I can fix them all, I have no idea it's a bad thing.
[1] https://lore.kernel.org/linux-fsdevel/20190830163910.GB29603@xxxxxxxxxxxxx/
[2] https://lore.kernel.org/linux-fsdevel/20190831064853.GA162401@architecture4/
Thanks,
Gao Xiang
