If we don't drop caches used in old offset, we can get old data from new offset, which gives unexpected data to user.
Martijn found a loopback bug in the below scenario. 1) LOOP_SET_FD loads first two pages on loop file 2) LOOP_SET_STATUS64 changes the offset on the loop file 3) mount is failed due to the cached pages having wrong superblock This patch drops caches when we change lo->offset. Cc: Jens Axboe <[email protected]> Cc: [email protected] Reported-by: Martijn Coenen <[email protected]> Signed-off-by: Jaegeuk Kim <[email protected]> --- drivers/block/loop.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/drivers/block/loop.c b/drivers/block/loop.c index cb0cc8685076..f073a3f1a7cd 100644 --- a/drivers/block/loop.c +++ b/drivers/block/loop.c @@ -1154,6 +1154,12 @@ loop_set_status(struct loop_device *lo, const struct loop_info64 *info) if (lo->lo_offset != info->lo_offset || lo->lo_sizelimit != info->lo_sizelimit) { + struct block_device *bdev = lo->lo_device; + + /* drop stale caches used in old offset */ + sync_blockdev(bdev); + kill_bdev(bdev); + if (figure_loop_size(lo, info->lo_offset, info->lo_sizelimit)) { err = -EFBIG; goto exit; -- 2.19.0.605.g01d371f741-goog
