On Sun, Jun 2, 2019 at 9:21 AM Yan, Zheng <zyan@xxxxxxxxxx> wrote:
>
> On 6/2/19 10:43 AM, Al Viro wrote:
> > On Sun, Jun 02, 2019 at 10:25:46AM +0800, Yan, Zheng wrote:
> >> remove_session_caps() relies on __wait_on_freeing_inode(), to wait for
> >> freezing inode to remove its caps. But VFS wakes freeing inode waiters
> >> before calling destroy_inode().
> >
> > *blink*
> >
> > Which tree is that against?
> >
> >> -static void ceph_i_callback(struct rcu_head *head)
> >> -{
> >> - struct inode *inode = container_of(head, struct inode, i_rcu);
> >> - struct ceph_inode_info *ci = ceph_inode(inode);
> >> -
> >> - kfree(ci->i_symlink);
> >> - kmem_cache_free(ceph_inode_cachep, ci);
> >> -}
> >
> > ... is gone from mainline, and AFAICS not reintroduced in ceph tree.
> > What am I missing here?
> >
>
> Sorry, I should send it ceph-devel list
Hi Zheng,
I have rebased testing on top of 5.2-rc6 and dropped this patch.
It conflicts with Al's cfa6d41263ca ("ceph: use ->free_inode()") in
mainline and I don't see a new version.
Please take a look.
Thanks,
Ilya
