On 2022/4/19 12:09, Andrew Morton wrote: > On Sat, 16 Apr 2022 16:14:16 +0800 Miaohe Lin <linmiaohe@xxxxxxxxxx> wrote: > >> We can't assume pte_offset_map_lock will return same orig_pte value. So >> it's necessary to reacquire the orig_pte or pte_unmap_unlock will unmap >> the stale pte. > > hm, where did you learn this info about pte_offset_map_lock()? > > I assume this is from code inspection only? No observed runtime failures? Yes, this is from code inspection. There is no observed runtime failures now due to the race window being really small. And this becomes noop in !CONFIG_HIGHMEM system (CONFIG_HIGHMEM system should be rare now). But this could be triggered theoretically. Thanks! > . >
