Re: [PATCH v2] mm/oom_kill: wake futex waiters before annihilating victim shared mutex

[Joel Savitz <jsavitz@redhat.com>]

> I am afraid we can't call futex_exit_release() under rcu_read_lock()
> because it might sleep.

Ah that's too bad. Is there an equivalent atomic call suitable for this purpose?

>
>  BUG: sleeping function called from invalid context at kernel/locking/mutex.c:577
>  in_atomic(): 1, irqs_disabled(): 0, non_block: 0, pid: 1509, name: lsbug
>  preempt_count: 1, expected: 0
>  3 locks held by lsbug/1509:
>   #0: ffff00004de99c98 (&mm->mmap_lock){++++}-{3:3}, at: do_page_fault
>   #1: ffff800010fd8308 (oom_lock){+.+.}-{3:3}, at: __alloc_pages_slowpath.constprop.0
>   __alloc_pages_may_oom at /usr/src/linux-next/mm/page_alloc.c:4278
>   (inlined by) __alloc_pages_slowpath at /usr/src/linux-next/mm/page_alloc.c:5058
>   #2: ffff000867b3b0c0 (&p->alloc_lock){+.+.}-{2:2}, at: find_lock_task_mm
>   find_lock_task_mm at /usr/src/linux-next/mm/oom_kill.c:145
>  CPU: 5 PID: 1509 Comm: lsbug Not tainted 5.16.0-rc5-next-20211214+ #172
>  Hardware name: MiTAC RAPTOR EV-883832-X3-0001/RAPTOR, BIOS 1.6 06/28/2020
>  Call trace:
>   dump_backtrace
>   show_stack
>   dump_stack_lvl
>   dump_stack
>   __might_resched
>   __might_sleep
>   __mutex_lock
>   mutex_lock_nested
>   futex_cleanup_begin
>   futex_cleanup_begin at /usr/src/linux-next/kernel/futex/core.c:1071
>   futex_exit_release
>   __oom_kill_process
>   oom_kill_process
>   out_of_memory
>   __alloc_pages_slowpath.constprop.0
>   __alloc_pages
>   alloc_pages_vma
>   alloc_zeroed_user_highpage_movable
>   do_anonymous_page
>   __handle_mm_fault
>   handle_mm_fault
>   do_page_fault
>   do_translation_fault
>   do_mem_abort
>   el0_da
>   el0t_64_sync_handler
>   el0t_64_sync
>  =============================
>  [ BUG: Invalid wait context ]
>  5.16.0-rc5-next-20211214+ #172 Tainted: G        W
>  -----------------------------
>  lsbug/1509 is trying to lock:
>  ffff000867b3ba98 (&tsk->futex_exit_mutex){+.+.}-{3:3}, at: futex_cleanup_begin
>  other info that might help us debug this:
>  context-{4:4}
>  3 locks held by lsbug/1509:
>   #0: ffff00004de99c98 (&mm->mmap_lock){++++}-{3:3}, at: do_page_fault
>   #1: ffff800010fd8308 (oom_lock){+.+.}-{3:3}, at: __alloc_pages_slowpath.constprop.0
>   #2: ffff000867b3b0c0 (&p->alloc_lock){+.+.}-{2:2}, at: find_lock_task_mm
>  stack backtrace:
>  CPU: 5 PID: 1509 Comm: lsbug Tainted: G        W         5.16.0-rc5-next-20211214+ #172
>  Hardware name: MiTAC RAPTOR EV-883832-X3-0001/RAPTOR, BIOS 1.6 06/28/2020
>  Call trace:
>   dump_backtrace
>   show_stack
>   dump_stack_lvl
>   dump_stack
>   __lock_acquire
>   lock_acquire
>   __mutex_lock
>   mutex_lock_nested
>   futex_cleanup_begin
>   futex_exit_release
>   __oom_kill_process
>   oom_kill_process
>   out_of_memory
>   __alloc_pages_slowpath.constprop.0
>   __alloc_pages
>   alloc_pages_vma
>   alloc_zeroed_user_highpage_movable
>   do_anonymous_page
>   __handle_mm_fault
>   handle_mm_fault
>   do_page_fault
>   do_translation_fault
>   do_mem_abort
>   el0_da
>   el0t_64_sync_handler
>   el0t_64_sync
>
> > ---
> >  mm/oom_kill.c | 12 ++++++++++++
> >  1 file changed, 12 insertions(+)
> >
> > diff --git a/mm/oom_kill.c b/mm/oom_kill.c
> > index 1ddabefcfb5a..884a5f15fd06 100644
> > --- a/mm/oom_kill.c
> > +++ b/mm/oom_kill.c
> > @@ -44,6 +44,7 @@
> >  #include <linux/kthread.h>
> >  #include <linux/init.h>
> >  #include <linux/mmu_notifier.h>
> > +#include <linux/futex.h>
> >
> >  #include <asm/tlb.h>
> >  #include "internal.h"
> > @@ -885,6 +886,11 @@ static void __oom_kill_process(struct task_struct *victim, const char *message)
> >       count_vm_event(OOM_KILL);
> >       memcg_memory_event_mm(mm, MEMCG_OOM_KILL);
> >
> > +     /*
> > +      * We call futex_exit_release() on the victim task to ensure any waiters on any
> > +      * process-shared futexes held by the victim task are woken up.
> > +      */
> > +     futex_exit_release(victim);
> >       /*
> >        * We should send SIGKILL before granting access to memory reserves
> >        * in order to prevent the OOM victim from depleting the memory
> > @@ -930,6 +936,12 @@ static void __oom_kill_process(struct task_struct *victim, const char *message)
> >                */
> >               if (unlikely(p->flags & PF_KTHREAD))
> >                       continue;
> > +             /*
> > +              * We call futex_exit_release() on any task p sharing the
> > +              * victim->mm to ensure any waiters on any
> > +              * process-shared futexes held by task p are woken up.
> > +              */
> > +             futex_exit_release(p);
> >               do_send_sig_info(SIGKILL, SEND_SIG_PRIV, p, PIDTYPE_TGID);
> >       }
> >       rcu_read_unlock();
> > --
> > 2.27.0
> >
>

Best,
Joel Savitz