linux-kernel - Re: Multiple oom_reaper BUGs: unmap_page_range racing with exit

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <alpine.DEB.2.10.1712080118420.145074@chino.kir.corp.google.com>
Date:   Fri, 8 Dec 2017 01:26:46 -0800 (PST)
From:   David Rientjes <rientjes@...gle.com>
To:     Michal Hocko <mhocko@...nel.org>
cc:     Tetsuo Handa <penguin-kernel@...ove.sakura.ne.jp>,
        akpm@...ux-foundation.org, aarcange@...hat.com,
        linux-kernel@...r.kernel.org, linux-mm@...ck.org
Subject: Re: Multiple oom_reaper BUGs: unmap_page_range racing with
 exit_mmap

On Thu, 7 Dec 2017, David Rientjes wrote:

> I'm backporting and testing the following patch against Linus's tree.  To 
> clarify an earlier point, we don't actually have any change from upstream 
> code that allows for free_pgtables() before the 
> set_bit(MMF_OOM_SKIP);down_write();up_write() cycle.
> 
> diff --git a/include/linux/oom.h b/include/linux/oom.h
> --- a/include/linux/oom.h
> +++ b/include/linux/oom.h
> @@ -66,6 +66,15 @@ static inline bool tsk_is_oom_victim(struct task_struct * tsk)
>  	return tsk->signal->oom_mm;
>  }
>  
> +/*
> + * Use this helper if tsk->mm != mm and the victim mm needs a special
> + * handling. This is guaranteed to stay true after once set.
> + */
> +static inline bool mm_is_oom_victim(struct mm_struct *mm)
> +{
> +	return test_bit(MMF_OOM_VICTIM, &mm->flags);
> +}
> +
>  /*
>   * Checks whether a page fault on the given mm is still reliable.
>   * This is no longer true if the oom reaper started to reap the
> diff --git a/include/linux/sched/coredump.h b/include/linux/sched/coredump.h
> --- a/include/linux/sched/coredump.h
> +++ b/include/linux/sched/coredump.h
> @@ -71,6 +71,7 @@ static inline int get_dumpable(struct mm_struct *mm)
>  #define MMF_HUGE_ZERO_PAGE	23      /* mm has ever used the global huge zero page */
>  #define MMF_DISABLE_THP		24	/* disable THP for all VMAs */
>  #define MMF_DISABLE_THP_MASK	(1 << MMF_DISABLE_THP)
> +#define MMF_OOM_VICTIM		25	/* mm is the oom victim */
>  
>  #define MMF_INIT_MASK		(MMF_DUMPABLE_MASK | MMF_DUMP_FILTER_MASK |\
>  				 MMF_DISABLE_THP_MASK)
> diff --git a/mm/mmap.c b/mm/mmap.c
> --- a/mm/mmap.c
> +++ b/mm/mmap.c
> @@ -3019,20 +3019,20 @@ void exit_mmap(struct mm_struct *mm)
>  	/* Use -1 here to ensure all VMAs in the mm are unmapped */
>  	unmap_vmas(&tlb, vma, 0, -1);
>  
> -	set_bit(MMF_OOM_SKIP, &mm->flags);
> -	if (unlikely(tsk_is_oom_victim(current))) {
> +	if (unlikely(mm_is_oom_victim(mm))) {
>  		/*
>  		 * Wait for oom_reap_task() to stop working on this
>  		 * mm. Because MMF_OOM_SKIP is already set before
>  		 * calling down_read(), oom_reap_task() will not run
>  		 * on this "mm" post up_write().
>  		 *
> -		 * tsk_is_oom_victim() cannot be set from under us
> +		 * mm_is_oom_victim() cannot be set from under us
>  		 * either because current->mm is already set to NULL
>  		 * under task_lock before calling mmput and oom_mm is
>  		 * set not NULL by the OOM killer only if current->mm
>  		 * is found not NULL while holding the task_lock.
>  		 */
> +		set_bit(MMF_OOM_SKIP, &mm->flags);
>  		down_write(&mm->mmap_sem);
>  		up_write(&mm->mmap_sem);
>  	}
> diff --git a/mm/oom_kill.c b/mm/oom_kill.c
> --- a/mm/oom_kill.c
> +++ b/mm/oom_kill.c
> @@ -683,8 +683,10 @@ static void mark_oom_victim(struct task_struct *tsk)
>  		return;
>  
>  	/* oom_mm is bound to the signal struct life time. */
> -	if (!cmpxchg(&tsk->signal->oom_mm, NULL, mm))
> +	if (!cmpxchg(&tsk->signal->oom_mm, NULL, mm)) {
>  		mmgrab(tsk->signal->oom_mm);
> +		set_bit(MMF_OOM_VICTIM, &mm->flags);
> +	}
>  
>  	/*
>  	 * Make sure that the task is woken up from uninterruptible sleep
> 

This passes all functional testing that I have and I can create a 
synthetic testcase that can trigger at least MMF_OOM_VICTIM getting set 
while oom_reaper is still working on an mm that this prevents, so feel 
free to add an

	Acked-by: David Rientjes <rientjes@...gle.com>

with a variant of your previous changelogs.  Thanks!

I think it would appropriate to cc stable for 4.14 and add a

Fixes: 212925802454 ("mm: oom: let oom_reap_task and exit_mmap run 
concurrently")

if nobody disagrees, which I think you may have already done on a previous 
iteration.

We can still discuss if there are any VM_LOCKED subtleties in the this 
thread, but I have no evidence that it is responsible for any issues.