dax: make sure inodes are flushed before destroy cache

[ Upstream commit a7e8de8 ]

A bug can be triggered by following command

$ modprobe nd_pmem && modprobe -r nd_pmem

[   10.060014] BUG dax_cache (Not tainted): Objects remaining in dax_cache on __kmem_cache_shutdown()
[   10.060938] Slab 0x0000000085b729ac objects=9 used=1 fp=0x000000004f5ae469 flags=0x200000000010200(slab|head|node)
[   10.062433] Call Trace:
[   10.062673]  dump_stack_lvl+0x34/0x44
[   10.062865]  slab_err+0x90/0xd0
[   10.063619]  __kmem_cache_shutdown+0x13b/0x2f0
[   10.063848]  kmem_cache_destroy+0x4a/0x110
[   10.064058]  __x64_sys_delete_module+0x265/0x300

This is caused by dax_fs_exit() not flushing inodes before destroy cache.
To fix this issue, call rcu_barrier() before destroy cache.

Signed-off-by: Tong Zhang <[email protected]>
Reviewed-by: Ira Weiny <[email protected]>
Reviewed-by: Christoph Hellwig <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Fixes: 7b6be84 ("dax: refactor dax-fs into a generic provider of 'struct dax_device' instances")
Signed-off-by: Dan Williams <[email protected]>
Signed-off-by: Sasha Levin <[email protected]>

drivers/dax/super.c

@@ -678,6 +678,7 @@ static int dax_fs_init(void)
 static void dax_fs_exit(void)
 {
 	kern_unmount(dax_mnt);
+	rcu_barrier();
 	kmem_cache_destroy(dax_cache);
 }