On Thu, 2019-03-21 at 20:08 +0500, Mikhail Gavrilov wrote:
> On Thu, 21 Mar 2019 at 18:21, Qian Cai <cai@xxxxxx> wrote:
> >
> > Does it come up with this page address every time?
> >
> > page:ffffcf49607ce000
>
> No it doesn't.
>
> $ journalctl | grep "page:"
> Mar 18 05:27:58 localhost.localdomain kernel: page:ffffdcd2607ce000 is
> uninitialized and poisoned
> Mar 20 22:29:19 localhost.localdomain kernel: page:ffffe4b7607ce000 is
> uninitialized and poisoned
> Mar 20 23:03:52 localhost.localdomain kernel: page:ffffd27aa07ce000 is
> uninitialized and poisoned
> Mar 21 09:29:29 localhost.localdomain kernel: page:ffffcf49607ce000 is
> uninitialized and poisoned
OK, those pages look similar enough. If you add this to __init_single_page() in
mm/page_alloc.c :
if (page == (void *)0xffffdcd2607ce000 || page == (void *)0xffffe4b7607ce000 ||
page == (void *)0xffffd27aa07ce000 || page == (void *)0xffffcf49607ce000) {
printk("KK page = %px\n", page);
dump_stack();
}
to see where those pages have been initialized in the first place.
