fs/dcache.c: fix spin lockup issue on nlru->lock
authorSahitya Tummala <stummala@codeaurora.org>
Mon, 10 Jul 2017 22:50:00 +0000 (15:50 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Mon, 10 Jul 2017 23:32:33 +0000 (16:32 -0700)
__list_lru_walk_one() acquires nlru spin lock (nlru->lock) for longer
duration if there are more number of items in the lru list.  As per the
current code, it can hold the spin lock for upto maximum UINT_MAX
entries at a time.  So if there are more number of items in the lru
list, then "BUG: spinlock lockup suspected" is observed in the below
path:

  spin_bug+0x90
  do_raw_spin_lock+0xfc
  _raw_spin_lock+0x28
  list_lru_add+0x28
  dput+0x1c8
  path_put+0x20
  terminate_walk+0x3c
  path_lookupat+0x100
  filename_lookup+0x6c
  user_path_at_empty+0x54
  SyS_faccessat+0xd0
  el0_svc_naked+0x24

This nlru->lock is acquired by another CPU in this path -

  d_lru_shrink_move+0x34
  dentry_lru_isolate_shrink+0x48
  __list_lru_walk_one.isra.10+0x94
  list_lru_walk_node+0x40
  shrink_dcache_sb+0x60
  do_remount_sb+0xbc
  do_emergency_remount+0xb0
  process_one_work+0x228
  worker_thread+0x2e0
  kthread+0xf4
  ret_from_fork+0x10

Fix this lockup by reducing the number of entries to be shrinked from
the lru list to 1024 at once.  Also, add cond_resched() before
processing the lru list again.

Link: http://marc.info/?t=149722864900001&r=1&w=2
Link: http://lkml.kernel.org/r/1498707575-2472-1-git-send-email-stummala@codeaurora.org
Signed-off-by: Sahitya Tummala <stummala@codeaurora.org>
Suggested-by: Jan Kara <jack@suse.cz>
Suggested-by: Vladimir Davydov <vdavydov.dev@gmail.com>
Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com>
Cc: Alexander Polakov <apolyakov@beget.ru>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
fs/dcache.c

index 7ece68d..6c30be6 100644 (file)
@@ -1160,11 +1160,12 @@ void shrink_dcache_sb(struct super_block *sb)
                LIST_HEAD(dispose);
 
                freed = list_lru_walk(&sb->s_dentry_lru,
-                       dentry_lru_isolate_shrink, &dispose, UINT_MAX);
+                       dentry_lru_isolate_shrink, &dispose, 1024);
 
                this_cpu_sub(nr_dentry_unused, freed);
                shrink_dentry_list(&dispose);
-       } while (freed > 0);
+               cond_resched();
+       } while (list_lru_count(&sb->s_dentry_lru) > 0);
 }
 EXPORT_SYMBOL(shrink_dcache_sb);