On (01/23/17 13:17), Minchan Kim wrote: > zram_reset_device waits ongoing writepage pages completed by > zram->refcount logic. However, it's pointless because before > the reset, we prevent further opening of zram by zram->claim > and flush all of pending IO by fsync_bdev so there should be > no pending IO at the zram_reset_device.
interesting.
looks reasonable to me.
> Signed-off-by: Minchan Kim <minc...@kernel.org>
Reviewed-by: Sergey Senozhatsky <sergey.senozhat...@gmail.com>
-ss
