Re: [tip: locking/urgent] futex: Handle transient "ownerless" rtmutex state correctly

From: Mike Galbraith
Date: Sat Nov 07 2020 - 12:06:13 EST


On Fri, 2020-11-06 at 21:26 +0000, tip-bot2 for Mike Galbraith wrote:
>
> ---
> kernel/futex.c | 16 ++++++++++++++--
> 1 file changed, 14 insertions(+), 2 deletions(-)
>
> diff --git a/kernel/futex.c b/kernel/futex.c
> index f8614ef..7406914 100644
> --- a/kernel/futex.c
> +++ b/kernel/futex.c
> @@ -2380,10 +2380,22 @@ retry:
> }
>
> /*
> - * Since we just failed the trylock; there must be an owner.
> + * The trylock just failed, so either there is an owner or
> + * there is a higher priority waiter than this one.
> */
> newowner = rt_mutex_owner(&pi_state->pi_mutex);
> - BUG_ON(!newowner);
> + /*
> + * If the higher priority waiter has not yet taken over the
> + * rtmutex then newowner is NULL. We can't return here with
> + * that state because it's inconsistent vs. the user space
> + * state. So drop the locks and try again. It's a valid
> + * situation and not any different from the other retry
> + * conditions.
> + */
> + if (unlikely(!newowner)) {
> + ret = -EAGAIN;
^^^

My box just discovered an unnoticed typo. That 'ret' should read 'err'
so we goto retry, else fbomb_v2 proggy will trigger gripeage.

[ 44.089233] fuse: init (API version 7.32)
[ 78.485163] ------------[ cut here ]------------
[ 78.485171] WARNING: CPU: 1 PID: 4557 at kernel/futex.c:2482 fixup_pi_state_owner.isra.17+0x125/0x350
[ 78.485171] ------------[ cut here ]------------
[ 78.485174] WARNING: CPU: 2 PID: 4559 at kernel/futex.c:1486 do_futex+0x920/0xaf0
<snip>

-Mike