Re: [PATCH 4/4] mm: swapoff: shmem_unuse() stop eviction without igrab()

[Hugh Dickins]

On Tue, 9 Apr 2019, Konstantin Khlebnikov wrote:
> On 08.04.2019 23:01, Hugh Dickins wrote:
> > -           if (!list_empty(&info->swaplist)) {
> > +           while (!list_empty(&info->swaplist)) {
> > +                   /* Wait while shmem_unuse() is scanning this inode...
> > */
> > +                   wait_var_event(&info->stop_eviction,
> > +                                  !atomic_read(&info->stop_eviction));
> >                     mutex_lock(&shmem_swaplist_mutex);
> >                     list_del_init(&info->swaplist);
> 
> Obviously, line above should be deleted.

Definitely. Worryingly stupid. I guess I left it behind while translating
from an earlier tree.  Many thanks for catching that in time, Konstantin.
I've rechecked the rest of this patch, and the others, and didn't find
anything else as stupid.

Andrew, please add this fixup for folding in - thanks:

[PATCH] mm: swapoff: shmem_unuse() stop eviction without igrab() fix

Fix my stupidity, thankfully caught by Konstantin.

Signed-off-by: Hugh Dickins <hu...@google.com>
---
Fix to fold into mm-swapoff-shmem_unuse-stop-eviction-without-igrab.patch

 mm/shmem.c |    1 -
 1 file changed, 1 deletion(-)

--- patch4/mm/shmem.c   2019-04-07 19:18:43.248639711 -0700
+++ patch5/mm/shmem.c   2019-04-09 11:24:32.745337734 -0700
@@ -1086,7 +1086,6 @@ static void shmem_evict_inode(struct ino
                        wait_var_event(&info->stop_eviction,
                                       !atomic_read(&info->stop_eviction));
                        mutex_lock(&shmem_swaplist_mutex);
-                       list_del_init(&info->swaplist);
                        /* ...but beware of the race if we peeked too early */
                        if (!atomic_read(&info->stop_eviction))
                                list_del_init(&info->swaplist);