Re: mm: swap: locking in release_pages()

[Michal Hocko <mhocko@suse.com>]

On Tue 05-04-22 16:00:54, Alexander Sverdlin wrote:
> Hello Michal!
> 
> On 05/04/2022 13:45, Michal Hocko wrote:
> >>>> 1. Crash of v5.4.170 on an ARM32 machine:
> >>>>
> >>>> Unable to handle kernel NULL pointer dereference at virtual address 00000104
> >>>> pgd = e138149d
> >>>> [00000104] *pgd=84d2fd003, *pmd=8ffd6f003
> >>>> Internal error: Oops: a07 [#1] PREEMPT SMP ARM
> >>>> ...
> >>>> CPU: 1 PID: 6172 Comm: AaSysInfoRColle Tainted: G    B      O      5.4.170-... #1
> >>>> Hardware name: Keystone
> >>>> PC is at release_pages+0x194/0x358
> >>>> LR is at release_pages+0x10c/0x358
> >>> Which LOC does this correspond to? (faddr2line should give you a nice
> >>> output).
> >> Sorry, I forgot this info in the initial report:
> >>
> >> this is indeed the del_page_from_lru_list() in this crash.
> > Could you be more specific please? Is the problem in list_del or
> > update_lru_size?
> 
> static inline void __list_del(struct list_head * prev, struct list_head * next)
> {
>         next->prev = prev;    <--

OK, I see. AFAICS this means that entry->next is NULL which doesn't look
like somebody else has done list_del as that would leave poison values
behind. Maybe somebody has clobbered the page state.

In any case I would recommend reproducing without stable patches and/or
with the current Linus tree.
-- 
Michal Hocko
SUSE Labs