It whould be better if put check_mem_permission after __get_free_page in
mem_write, to be same as function mem_read.
Hugh Dickins explained the reason.
check_mem_permission gets a reference to the mm. If we __get_free_page
after check_mem_permission, imagine what happens if the system is out
of memory, and the mm we're looking at is selected for killing by the
OOM killer: while we wait in __get_free_page for more memory, no memory
is freed from the selected mm because it cannot reach exit_mmap while
we hold that reference.
Reported-by: Jovi Zhang <bookjovi@gmail.com>
Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Acked-by: Hugh Dickins <hughd@google.com>
Reviewed-by: Stephen Wilson <wilsons@start.ca>
Cc: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
if (!task)
goto out_no_task;
+ copied = -ENOMEM;
+ page = (char *)__get_free_page(GFP_TEMPORARY);
+ if (!page)
+ goto out_task;
+
mm = check_mem_permission(task);
copied = PTR_ERR(mm);
if (IS_ERR(mm))
- goto out_task;
+ goto out_free;
copied = -EIO;
if (file->private_data != (void *)((long)current->self_exec_id))
goto out_mm;
- copied = -ENOMEM;
- page = (char *)__get_free_page(GFP_TEMPORARY);
- if (!page)
- goto out_mm;
-
copied = 0;
while (count > 0) {
int this_len, retval;
count -= retval;
}
*ppos = dst;
- free_page((unsigned long) page);
+
out_mm:
mmput(mm);
+out_free:
+ free_page((unsigned long) page);
out_task:
put_task_struct(task);
out_no_task: